Tag: Machine Learning

A3 in customer experience: Possibilities for personalisation

The value of A3 in customer experience

This report considers the financial value to a telco of using A3 technologies (analytics, automation and AI) to improve customer experience. It examines the key area which underpins much of this financial value – customer support channels – considering the trends in this area and how the area might change in future, shaping the requirement for A3.

Calculating the value of improving customer experience is complex: it can be difficult to identify the specific action that improved a customer’s perception of their experience, and then to assess the impact of this improvement on their subsequent behaviour.

While it is difficult to draw causal links between telcos’ A3 activities and customer perceptions and behaviours, there are still some clearly measurable financial benefits from these investments. We estimate this value by leveraging our broader analysis of the financial value of A3 in telecoms, and then zooming in on the specific pockets of value which relate to improved customer experience (e.g. churn reduction).

Enter your details below to request an extract of the report


 

The diagram below illustrates that there are two parts of the customer journey where A3 will add most value to customer experience:

  1. The performance of the network, services, devices and applications is increasingly dependent on automation and intelligence, with the introduction of 5G and cloud-native operations. Without A3 capabilities it will be difficult to meet quality of service standards, understand customer-affecting issues and turn up new services at speed.
  2. The contact centre remains one of the largest influencers of customer experience and one of the biggest users of automation, with the digital channels increasing in importance during the pandemic. Understanding the customer and the agent’s needs and providing information about issues the customer is experiencing to both parties are areas where more A3 should be used in future.

Where is the financial benefit of adding A3 within a typical telco customer journey?

A3 customer experience

Source: STL Partners, Charlotte Patrick Consult

As per this diagram, many of the most valuable uses for A3 are in the contact centre and digital channels. Improvements in customer experience will be tied with trends in both. These priority trends and potential A3 solutions are outlined the following two tables:
• The first shows contact centre priorities,
• The second shows priorities for the digital channels.

Priorities in the contact centre

A3 Contact centre

Priorities in the digital channel

A3 Digital channel

Table of Contents

  • Executive Summary
  • The value of A3 in customer experience
  • Use of A3 to improve customer experience
  • The most important uses of A3 for improving the customer experience
    • Complex data
    • Personalisation
    • Planning
    • Human-machine interaction
    • AI point solution
  • Conclusion
  • Appendix: Methodology for calculating financial value
  • Index

Related Research:

Enter your details below to request an extract of the report


 

Innovation leader case study: Telefónica Tech AI of Things

The origins of Telefónica Tech AI of Things

Telefónica LUCA was set up in 2016 to “enable corporate clients to understand their data and encourage a transparent and responsible use of that data”.

Before the creation of LUCA, Telefónica’s focus had been on developing assets and making acquisitions (e.g. Synergic Partners) to build strong internal capabilities around data and analytics – with some data monetisation capabilities housed within their Telefónica Digital unit (a global business unit selling products beyond connectivity, which was disbanded in 2016). Typical projects the team undertook related to using network data to make better decisioning for the network and marketing teams, and providing Telefónica Digital with external monetisation opportunities such as Smart Steps (aggregated, anonymised data for creation of vertical products) and Smart Digits (provision of consent-based data to the advertising industry).

Enter your details below to request an extract of the report


 

Creating the autonomous LUCA unit made a statement that Telefónica was serious about its strategy to offer data products to enterprise customers. Quoting from the original press release, “LUCA offered three lines of products and services:

The Business Insights area brings the value of anonymous and aggregated data on Telefónica’s networks for a wide range of clients. This includes Smart Steps, which is focused on mobility analysis solutions for more efficient planning. For example, to optimise transport networks and tourist management in cities, or in the case of a health emergency, in helping to better understand population movements and in limiting the spread of pandemics.

The analytical and external consultancy services for national and international clients will be provided by Synergic Partners, a company specialized in Big Data and Data Science which was acquired by Telefónica at the end of 2015.

Furthermore, LUCA will help its clients by providing BDaaS (Big Data as a Service) to empower clients to get the most out of their own data, using the Telefónica cloud infrastructure.”

The following table shows a timeline from the origins of LUCA in the Telefónica Digital business unit through to its merger into the Telefónica Tech AI of Things business in 2019 – illustrating the progression of its products and other major activities.

Timeline of Telefónica’s data monetisation business

Telefonica-data-monetisation-luca-AI-IoT

Source: STL Partners, Charlotte Patrick Consult

Points to note on the timeline above:

  • Telefónica stood out from its peers with the purchase of Synergic Partners in 2015 (bringing in 120 consultancy headcount). This provided not only another leg to the business with consulting capabilities, but also additional headcount to scope and sell their existing product sets.
  • Looking at the timeline, it took Telefónica two years from this purchase and the establishment LUCA to expand its portfolio. In 2018, a range of new, mainly IoT-related capabilities, were launched, built up from existing projects with individual customers.
  • Telefónica has added machine learning to its products across the timeframe, but in 2019 the development of NLP capability for use in Telefónica’s existing products, and an internal data science platform, were then productised for customers (see below discussion about its Aura product set).
  • As the number of products has expanded, the number of partnerships has also expanded, bringing specific platforms and capabilities which can be combined with Telefónica’s own data capabilities to provide added value (examples include CARTO which creates geographic visualisations of Telefónica’s data).
  • Looking at changing vertical priorities:
    • Telefónica has always been strong in the advertising sector, starting with products from O2 UK in 2012. The exact nature of what it has offered has changed over time and some capabilities have been sold, however, it still has a strong mobile marketing business and expects it data to become of more interest to brands/media agencies as the use of cookies diminishes across the next few years.
    • The retail sector offers opportunity, but has been challenging to target over the years. Although Telefónica has interesting data for retail companies, creating replicable products is challenging as the large retailers each have differing requirements and working with small cell data in-store can be expensive. The product set is therefore currently being simplified, as the pandemic has also reduced demand from retailers.

One of Telefónica’s key capabilities which is not clearly displayed in the timeline is the provision of services to the marketing teams of the various verticals it targets. These include analytics products which Telefónica has developed from its internal capabilities and other functionality such as pricing tools.

The formation of Telefónica Tech

In 2019, Telefónica LUCA became part of the newly formed, autonomous Telefónica Tech business unit. The organisation is split into two business areas: cybersecurity & cloud, and the assets from Telefónica LUCA combined with the IoT unit. The goal of Telefónica Tech is to:

  • Enable the financial markets to clearly see revenue progression. Telefónica’s stated aim is for sustained double digit growth, which it achieved with year-on-year growth of 13.6% in 2020, although the IoT and Big Data segment only grew 0.8% y-o-y in 2020, due to the impact of COVID-19 on IoT deployments, especially in retail. Showing signs of recovery, in H121 revenue growth in the IoT and Big Data segment rose to 8.1% y-o-y, and to 26% y-o-y for the whole of Telefónica Tech.
  • Coordinate innovation, particularly around post-pandemic opportunities such as remote working, e-health, e-commerce and digital transformation
  • Take advantage of global synergies and leveraging existing assets
  • Ease M&A and partnerships activity (it already has 300 partners to better reach new markets, including relations with 60 start-ups across products)
  • Build relationships with cloud providers (it has existing relationships with Microsoft, Google and SAP).

To better leverage existing assets, Telefónica LUCA was integrated with Telefónica’s IoT capabilities to create a more unified set of capabilities:

  1. IoT is seen as an enabling opportunity for AI, which can bring added value to Telefónica’s 10,000 IoT customers (with 35 million live IoT SIMs worldwide). Opportunities include provision of intelligence around “things” (for example, products to analyse sensor data) and then the addition of Business Insight services (i.e. analysis of aggregated, anonymised Telefónica data which adds further insight alongside the data coming from IoT devices).
  2. AI is now often a commodity discussion with C-Level prospects and Telefónica wishes to be seen as a strategic partner. Telefónica’s AI of Things proposition offers an execution layer and integration experts with security-by-design capabilities.
  3. Combining capabilities provides sales teams with an end-to-end value proposition, as the addition of AI is often complimentary to cloud transformation projects and the implementation of digital platforms.

There is a growing ecosystem in IoT and data which will generate more opportunities as both IoT solutions and ML/AI solutions mature, although it is not a straightforward decision for Telefónica on how to compete within this ecosystem.

Table of contents

  • Executive Summary
    • How successful has Telefónica been in data monetisation?
    • Learnings from Telefónica’s experience
    • Key success factors
    • Telefónica’s future strategy
  • Introduction
    • The origins of Telefónica Tech AI of Things
    • The formation of Telefónica Tech
  • Vision, mission and strategy
    • Scaling the business
    • Building a product set
    • Learnings from Telefónica Tech AI of Things
  • Organisational strategy
    • Where should the data monetisation team live?
    • Structure of Telefónica Tech AI of Things Team
    • External partnerships
    • Future plans
  • Data portfolio strategy
    • Tools and infrastructure
    • AI Suite
    • Vertical strategy
    • Product development beyond analytics
  • Conclusion and future moves

Related research

Enter your details below to request an extract of the report


 

A3 for enterprise: Where should telcos focus?

A3 capabilities operators can offer enterprise customers

In this research we explore the potential enterprise solutions leveraging analytics, AI and automation (A3) that telcos can offer their enterprise customers. Our research builds on a previous STL Partners report Telco data monetisation: What’s it worth? which modelled the financial opportunity for telco data monetisation – i.e. purely the machine learning (ML) and analytics component of A3 – for 200+ use cases across 13 verticals.

In this report, we expand our analysis to include the importance of different types of AI and automation in implementing the 200+ use cases for enterprises and assess the feasibility for telcos to acquire and integrate those capabilities into their enterprise services.

We identified eight different types of A3 capabilities required to implement our 200+ use cases.

These capability types are organised below roughly in order of the number of use cases for which they are relevant (i.e. people analytics is required in the most use cases, and human learning is needed in the fewest).

The ninth category, Data provision, does not actually require any AI or automation skills beyond ML for data management, so we include it in the list primarily because it remains an opportunity for telcos that do not develop additional A3 capabilities for enterprise.

Enter your details below to request an extract of the report


 

Most relevant A3 capabilities across 200+ use cases

9-types-of-A3-analytics-AI-automation

Most relevant A3 capabilities for leveraging enterprise solutions

People analytics: This is the strongest opportunity for telcos as it uses their comprehensive customer data. Analytics and machine learning are required for segmentation and personalisation of messaging or action. Any telco with a statistically-relevant market share can create products – although specialist sales capabilities are still essential.

IoT analytics: Although telcos offering IoT products do not immediately have access to the payload data from devices, the largest telcos are offering a range of products which use analytics/ML to detect patterns or spot anomalies from connected sensors and other devices.

Other analytics: Similar to IoT, the majority of other analytics A3 use cases are around pattern or anomaly detection, where integration of telco data can increase the accuracy and success of A3 solutions. Many of the use cases here are very specific to the vertical. For example, risk management in financial services or tracking of electronic prescriptions in healthcare – which means that a telco will need to have existing products and sales capability in these verticals to make it worthwhile adding in new analytics or ML capabilities.

Real time: These use cases mainly need A3 to understand and act on triggers coming from customer behaviour and have mixed appeal to telcos. Telcos already play a significant role in a small number of uses cases, such as mobile marketing. Some telcos are also active in less mature use cases such as patient messaging in healthcare settings (e.g. real-time reminders to take medication or remote monitoring of vulnerable adults). Of the rest of the use cases that require real time automation, a subset could be enhanced with messaging. This would primarily be attractive to mobile operators, especially if they offer broader relevant enterprise solutions – for example, if a telco was involved in a connected public transport solution, then it could also offer passenger messaging.

Remote monitoring/control: Solutions track both things and people and use A3 to spot issues, do diagnostic analysis and prescribe solutions to the problems identified. The larger telcos already have solutions in some verticals, and 5G may bring more opportunities, such as monitoring of remote sites or traffic congestion monitoring.

Video analytics: Where telcos have CCTV implementations or video, there is opportunity to add in analytics solutions (potentially at the edge).

Human interactions: The majority of telco opportunities here relate to the provision of chatbots into enterprise contact centres.

Human learning: A group of low feasibility use cases around training (for example, an engineer on a manufacturing floor who uses a heads-up augmented/virtual reality (AR/VR) display to understand the resolution to a problem in front of them) or information provision (for example, providing retail customers with information via AR applications).

 

Table of Contents

  • Executive Summary
    • Which A3 capabilities should telcos prioritise?
    • What makes an investment worthwhile?
    • Next steps
  • Introduction
  • Vertical opportunities
    • Key takeaways
  • A3 technology: Where should telcos focus?
    • Key takeaways
    • Assessing the telco opportunity for nine A3 capabilities
  • Verizon case study
  • Details of vertical opportunities
  • Conclusion
  • Appendix 1 – full list of 200 use cases

 

Enter your details below to request an extract of the report


Are telcos smart enough to make money work?

Telco consumer financial services propositions

Telcos face a perplexing challenge in consumer markets. On the one hand, telcos’ standing with consumers has improved through the COVID-19 pandemic, and demand for connectivity is strong and continues to grow. On the other hand, most consumers are not spending more money with telcos because operators have yet to create compelling new propositions that they can charge more for. In the broadest sense, telcos need to (and can in our view) create more value for consumers and society more generally.

Download the report extract

As discussed in our previous research, we believe the world is now entering a “Coordination Age” in which multiple stakeholders will work together to maximize the potential of the planet’s natural and human resources. New technologies – 5G, analytics, AI, automation, cloud – are making it feasible to coordinate and optimise the allocation of resources in real-time. As providers of connectivity that generates vast amounts of relevant data, telcos can play an important role in enabling this coordination. Although some operators have found it difficult to expand beyond connectivity, the opportunity still exists and may actually be expanding.

In this report, we consider how telcos can support more efficient allocation of capital by playing in the financial services sector.  Financial services (banking) sits in a “sweet spot” for operators: economies of scale are available at a national level, connected technology can change the industry.

Financial Services in the Telecoms sweet spot

financial services

Source STL Partners

The financial services industry is undergoing major disruption brought about by a combination of digitisation and liberalisation – new legislation, such as the EU’s Payment Services Directive, is making it easier for new players to enter the banking market. And there is more disruption to come with the advent of digital currencies – China and the EU have both indicated that they will launch digital currencies, while the U.S. is mulling going down the same route.

A digital currency is intended to be a digital version of cash that is underpinned directly by the country’s central bank. Rather than owning notes or coins, you would own a deposit directly with the central bank. The idea is that a digital currency, in an increasingly cash-free society, would help ensure financial stability by enabling people to store at least some of their money with a trusted official platform, rather than a company or bank that might go bust. A digital currency could also make it easier to bring unbanked citizens (the majority of the world’s population) into the financial system, as central banks could issue digital currencies directly to individuals without them needing to have a commercial bank account. Telcos (and other online service providers) could help consumers to hold digital currency directly with a central bank.

Although the financial services industry has already experienced major upheaval, there is much more to come. “There’s no question that digital currencies and the underlying technology have the potential to drive the next wave in financial services,” Dan Schulman, the CEO of PayPal told investors in February 2021. “I think those technologies can help solve some of the fundamental problems of the system. The fact that there’s this huge prevalence and cost of cash, that there’s lack of access for so many parts of the population into the system, that there’s limited liquidity, there’s high friction in commerce and payments.”

In light of this ongoing disruption, this report reviews the efforts of various operators, such as Orange, Telefónica and Turkcell, to expand into consumer financial services, notably the provision of loans and insurance. A close analysis of their various initiatives offers pointers to the success criteria in this market, while also highlighting some of the potential pitfalls to avoid.

Table of contents

  • Executive Summary
  • Introduction
  • Potential business models
    • Who are you serving?
    • What are you doing for the people you serve?
    • M-Pesa – a springboard into an array of services
    • Docomo demonstrates what can be done
    • But the competition is fierce
  • Applying AI to lending and insurance
    • Analysing hundreds of data points
    • Upstart – one of the frontrunners in automated lending
    • Takeaways
  • From payments to financial portal
    • Takeaways
  • Turkcell goes broad and deep
    • Paycell has a foothold
    • Consumer finance takes a hit
    • Regulation moving in the right direction
    • Turkcell’s broader expansion plans
    • Takeaways
  • Telefónica targets quick loans
    • Growing competition
    • Elsewhere in Latin America
    • Takeaways
  • Momentum builds for Orange
    • The cost of Orange Bank
    • Takeaways
  • Conclusions and recommendations
  • Index

This report builds on earlier STL Partners research, including:

Download the report extract

Commerce and connectivity: A match made in heaven?

Rakuten and Reliance: The exceptions or the rule?

Over the past decade, STL Partners has analysed how connectivity, commerce and content have become increasingly interdependent – as both shopping and entertainment go digital, telecoms networks have become key distribution channels for all kinds of consumer businesses. Equally, the growing availability of digital commerce and content are driving demand for connectivity both inside and outside the home.

To date, the top tier of consumer Internet players – Google, Apple, Amazon, Alibaba, Tencent and Facebook – have tended to focus on trying to dominate commerce and content, largely leaving the provision of connectivity to the conventional telecoms sector. But now some major players in the commerce market, such as Rakuten in Japan and Reliance in India, are pushing into connectivity, as well as content.

This report considers whether Rakuten’s and Reliance’s efforts to combine content, commerce and connectivity into a single package is a harbinger of things to come or the exceptions that will prove the longstanding rule that telecoms is a distinct activity with few synergies with adjacent sectors. The provision of connectivity has generally been regarded as a horizontal enabler for other forms of economic activity, rather than part of a vertically-integrated service stack.

This report also explores the extent to which new technologies, such as cloud-native networks and open radio access networks, and an increase in licence-exempt spectrum, are making it easier for companies in adjacent sectors to provide connectivity. Two chapters cover Google and Amazon’s connectivity strategies respectively, analysing the moves they have made to date and what they may do in future. The final section of this report draws some conclusions and then considers the implications for telcos.

This report builds on earlier STL Partners research, including:

Enter your details below to download an extract of the report


Mixing commerce and connectivity

Over the past decade, the smartphone has become an everyday shopping tool for billions of people, particularly in Asia. As a result, the smartphone display has become an important piece of real estate for the global players competing for supremacy in the digital commerce market. That real estate can be accessed via a number of avenues – through the handset’s operating system, a web browser, mobile app stores or through the connectivity layer itself.

As Google and Apple exercise a high degree of control over smartphone operating systems, popular web browsers and mobile app stores, other big digital commerce players, such as Amazon, Facebook and Walmart, risk being marginalised. One way to avoid that fate may be to play a bigger role in the provision of wireless connectivity as Reliance Industries is doing in India and Rakuten is doing in Japan.

For telcos, this is potentially a worrisome prospect. By rolling out its own greenfield mobile network, e-commerce, and financial services platform Rakuten has brought disruption and low prices to Japan’s mobile connectivity market, putting pressure on the incumbent operators. There is a clear danger that digital commerce platforms use the provision of mobile connectivity as a loss leader to drive to traffic to their other services.

Table of Contents

  • Executive Summary
  • Introduction
  • Mixing connectivity and commerce
    • Why Rakuten became a mobile network operator
    • Will Rakuten succeed in connectivity?
    • Why hasn’t Rakuten Mobile broken through?
    • Borrowing from the Amazon playbook
    • How will the hyperscalers react?
  • New technologies, new opportunities
    • Capacity expansion
    • Unlicensed and shared spectrum
    • Cloud-native networks and Open RAN attract new suppliers
    • Reprogrammable SIM cards
  • Google: Knee deep in connectivity waters
    • Google Fiber and Fi maintain a holding pattern
    • Google ramps up and ramps down public Wi-Fi
    • Google moves closer to (some) telcos
    • Google Cloud targets telcos
    • Big commitment to submarine/long distance infrastructure
    • Key takeaways: Vertical optimisation not integration
  • Amazon: A toe in the water
    • Amazon Sidewalk
    • Amazon and CBRS
    • Amazon’s long distance infrastructure
    • Takeaways: Control over connectivity has its attractions
  • Conclusions and implications for telcos in digital commerce/content
  • Index

Enter your details below to download an extract of the report


Telco A3: Skilling up for the long term

Telcos must master automation, analytics and AI (A3) skills to remain competitive

A3 will permeate all aspects of telcos’ and their customers’ operations, improving efficiency, customer experience, and the speed of innovation. Therefore, whether a telecoms operator is focused on its core connectivity business, or seeking to build new value beyond connectivity, developing widespread understanding of value of A3 and disseminating fundamental automation and AI skills across the organisation should be a core strategic goal. Our surveys on industry priorities suggest that operators recognise this need, and automation and AI are correspondingly rising up the agenda.

Expected technology priority change by organisation type, May 2020

technology investment priorities telecoms May 2020

*Updated January 2021 survey results will be published soon. Source: STL Partners survey, 222 respondents.

Enter your details below to request an extract of the report


Key findings on operators’ A3 strategies

Based on deep dive interviews with 8 telcos, as well as insights from 8 more telcos gathered from previous research programmes.

  • Less advanced telcos are creating a set of basic structures and procedures, as well as beginning to develop a single view of the customer
  • Having a single version of the truth appears to be an ongoing issue for all – alongside continued work on data quality
  • As full end-to-end automation is not a realistic goal for the next few years, interviewees were seeking to prioritise the right journeys to be automated in the short term
  • Reskilling and education of staff was an area of importance for many but not all
  • Just one company had less ambitious data-related aims due to the specialist nature of their services and smaller size of the company – saying that they worked with data on an as-needed basis and had no plans to develop dedicated data science headcount

Preparing for the future: There are four areas where A3 will impact telcos’ businesses

four A3 areas impacting telcos

Source: Charlotte Patrick Consult, STL Partners

In this report we outline the skills and capabilities telcos will need in order to navigate these changes. We break out these skills into four layers:

  1. The basic skillset: What operators need to remain competitive over the short term
  2. The next 5 years: The skills virtually all telcos will need to build or acquire to remain competitive in the medium term (exceptions include small or specialist telcos, or those in less competitive markets)
  3. The next 10 years: The skills and organisational changes telcos will need to achieve within a 10 year timeframe to remain competitive in the long term
  4. Beyond connectivity (5–10 year horizon): This includes A3 skills that telcos will need to be successful strategic partners for customers and suppliers, and to thrive in ecosystem business models. These will be essential for telcos seeking to play a coordination role in IoT, edge, or industry ecosystems.

Table of contents

  • Executive Summary
  • Telcos’ current strategic direction
    • Deep dive into 8 operator strategies
    • Overview of 8 more operator strategies
  • How A3 technologies are evolving
    • Deep dive into 40 A3 applications that will impact telcos’ businesses
    • Internal capabilities
    • Customer requirements
    • Technology changes
    • Organisational change
  • A timeline of telco A3 skills evolution
    • The basic skillset
    • The next 5 years
    • The next 10 years
    • Beyond connectivity

Enter your details below to request an extract of the report


Fighting the fakes: How telcos can help

Internet platforms need a frictionless solution to fight the fakes

On the Internet, the old adage, nobody knows you are a dog, can still ring true. All of the major Internet platforms, with the partial exception of Apple, are fighting frauds and fakes. That’s generally because these platforms either allow users to remain anonymous or because they use lax authentication systems that prioritise ease-of-use over rigour. Some people then use the cloak of anonymity in many different ways, such as writing glowing reviews of products they have never used on Amazon (in return for a payment) or enthusiastic reviews of restaurants owned by friends on Tripadvisor. Even the platforms that require users to register financial details are open to abuse. There have been reports of multiple scams on eBay, while regulators have alleged there has been widespread sharing of Uber accounts among drivers in London and other cities.

At the same time, Facebook/WhatsApp, Google/YouTube, Twitter and other social media services are experiencing a deluge of fake news, some of which can be very damaging for society. There has been a mountain of misinformation relating to COVID-19 circulating on social media, such as the notion that if you can hold your breath for 10 seconds, you don’t have the virus. Fake news is alleged to have distorted the outcome of the U.S. presidential election and the Brexit referendum in the U.K.

In essence, the popularity of the major Internet platforms has made them a target for unscrupulous people who want to propagate their world views, promote their products and services, discredit rivals and have ulterior (and potentially criminal) motives for participating in the gig economy.

Although all the leading Internet platforms use tools and reporting mechanisms to combat misuse, they are still beset with problems. In reality, these platforms are walking a tightrope – if they make authentication procedures too cumbersome, they risk losing users to rival platforms, while also incurring additional costs. But if they allow a free-for-all in which anonymity reigns, they risk a major loss of trust in their services.

In STL Partners’ view, the best way to walk this tightrope is to use invisible authentication – the background monitoring of behavioural data to detect suspicious activities. In other words, you keep the Internet platform very open and easy-to-use, but algorithms process the incoming data and learn to detect the patterns that signal potential frauds or fakes. If this idea were taken to an extreme, online interactions and transactions could become completely frictionless. Rather than asking a person to enter a username and password to access a service, they can be identified through the device they are using, their location, the pattern of keystrokes and which features they access once they are logged in. However, the effectiveness of such systems depends heavily on the quality and quantity of data they are feeding on.

In come telcos

This report explores how telcos could use their existing systems and data to help the major Internet companies to build better systems to protect the integrity of their platforms.

It also considers the extent to which telcos will need to work together to effectively fight fraud, just as they do to combat telecoms-related fraud and prevent stolen phones from being used across networks. For most use cases, the telcos in each national market will generally need to provide a common gateway through which a third party could check attributes of the user of a specific mobile phone number. As they plot their way out of the current pandemic, governments are increasingly likely to call for such gateways to help them track the spread of COVID-19 and identify people who may have become infected.

Request a report extract

Using big data to combat fraud

In the financial services sector, artificial intelligence (AI) is now widely used to help detect potentially fraudulent financial transactions. Learning from real-world examples, neural networks can detect the behavioural patterns associated with fraud and how they are changing over time. They can then create a dynamic set of thresholds that can be used to trigger alarms, which could prompt a bank to decline a transaction.

In a white paper published in 2019, IBM claimed its AI and cognitive solutions are having a major impact on transaction monitoring and payment fraud modelling. In one of several case studies, the paper describes how the National Payment Switch in France (STET) is using behavioural information to reduce fraud losses by US$100 million annually. Owned by a consortium of financial institutions, STET processes more than 30 billion credit and debit card, cross-border, domestic and on-us payments annually.

STET now assesses the fraud risk for every authorisation request in real time. The white paper says IBM’s Safer Payments system generates a risk score, which is then passed to banks, issuers and acquirers, which combine it with customer information to make a decision on whether to clear or decline the transaction. IBM claims the system can process up to 1,200 transactions per second, and can compute a risk score in less than 10 milliseconds. While STET itself doesn’t have any customer data or data from other payment channels, the IBM system looks across all transactions, countrywide, as well as creating “deep behavioural profiles for millions of cards and merchants.”

Telcos, or at least the connectivity they provide, are also helping banks combat fraud. If they think a transaction is suspicious, banks will increasingly send a text message or call a customer’s phone to check whether they have actually initiated the transaction. Now, some telcos, such as O2 in the UK, are making this process more robust by enabling banks to check whether the user’s SIM card has been swapped between devices recently or if any call diverts are active – criminals sometimes pose as a specific customer to request a new SIM. All calls and texts to the number are then routed to the SIM in the fraudster’s control, enabling them to activate codes or authorisations needed for online bank transfers, such as a one-time PINs or passwords.

As described below, this is one of the use cases supported by Mobile Connect, a specification developed by the GSMA, to enable mobile operators to take a consistent approach to providing third parties with identification, authentication and attribute-sharing services. The idea behind Mobile Connect is that a third party, such as a bank, can access these services regardless of which operator their customer subscribes to.

Adapting telco authentication for Amazon, Uber and Airbnb

Telcos could also provide Internet platforms, such as Amazon, Uber and Airbnb, with identification, authentication and attribute-sharing services that will help to shore up trust in their services. Building on their nascent anti-fraud offerings for the financial services industry, telcos could act as intermediaries, authenticating specific attributes of an individual without actually sharing personal data with the platform.

STL Partners has identified four broad data sets telcos could use to help combat fraud:

  1. Account activity – checking which individual owns which SIM card and that the SIM hasn’t been swapped recently;
  2. Movement patterns – tracking where people are and where they travel frequently to help identify if they are who they say they are;
  3. Contact patterns – establishing which individuals come into contact with each other regularly;
  4. Spending patterns – monitoring how much money an individual spends on telecoms services.

Table of contents

  • Executive Summary
  • Introduction
  • Using big data to combat fraud
    • Account activity
    • Movement patterns
    • Contact patterns
    • Spending patterns
    • Caveats and considerations
  • Limited progress so far
    • Patchy adoption of Mobile Connect
    • Mobile identification in the UK
    • Turkcell employs machine learning
  • Big Internet use cases
    • Amazon – grappling with fake product reviews
    • Facebook and eBay – also need to clampdown
    • Google Maps and Tripadvisor – targets for fake reviews
    • Uber – serious safety concerns
    • Airbnb – balancing the interests of hosts and guests
  • Conclusions
  • Index

Request STL research insights overview pack

End-to-end network automation: Why and how to do it?

Automation, analytics and AI: A3 unlocks value for operators

STL Partners has been writing about automation, artificial intelligence (AI) and data analytics for several years. While the three have overlapping capabilities and often a single use case will rely upon a combination, they are also distinct in their technical outcomes.

Distinctions between the three As

Source: STL Partners

Operators have been heavily investing in A3 use cases for several years and are making significant progress. Efforts can be broadly broken down into five different domains: sales and marketing, customer experience, network planning and operations, service innovation and other operations. Some of these domains, such as sales and marketing and customer experience, are more mature, with significant numbers of use cases moving beyond R&D and PoCs into live, scaled deployments. In comparison, other domains, like service innovation, are typically less mature, despite the potential new revenue opportunities attached to them.

Five A3 use case domains

Source: STL Partners

Use cases often overlap across domains. For example, a Western European operator has implemented an advanced analytics platform that monitors network performance, and outputs a unique KPI that, at a per subscriber level, indicates the customer experience of the network. This can be used to trigger an automated marketing campaign to customers who are experiencing issues with their network performance (e.g. an offer for free mobile hotspot until issues are sorted). In this way, it spans both customer experience and network operations. For the purpose of this paper, however, we will primarily focus on automation use cases in the network domain.  We have modelled the financial value of A3 for telcos: Mapping the financial value.

Request a report extract

The time is ripe for network automation now

Network automation is not new. In fact, it’s been a core part of operator’s network capabilities since Almon Strowger invented the Strowger switch (in 1889), automating the process of the telephone exchange. Anecdotally, Strowger (an undertaker by profession) came up with this invention because the wife of a rival funeral parlour owner, working at the local community switchboard, was redirecting customers calling for Strowger to her own husband’s business.

Early advertising called the Strowger switch the “girl-less, cuss-less, out-of-order-less, wait-less telephone” or, in other words, free from human error and faster than the manual switchboard system. While this example is more than 100 years old, many of the benefits of automation that it achieved are still true today; automation can provide operators with the ability to deliver services on-demand, without the wait, and free from human error (or worse still, malevolent intent).

Despite automation not being a new phenomenon, STL Partners has identified six key reasons why network automation is something operators should prioritise now:

  • Only with automation can operators deliver the degree of agility that customers will demand. Customers today expect the kind of speed, accuracy and flexibility of service that can only be achieved in a cost-effective manner with high degrees of network automation. This can be both consumer customers (e.g. for next generation network services like VR/AR applications, gaming, high-definition video streaming etc.) or enterprise customers (e.g. for creating a network slice that is spun up for a weekend for a specific big event). With networks becoming increasingly customised, operators must automate their systems (across both OSS and BSS) to ensure that they can deliver these services without a drastic increase in their operating costs.
    One  wholesale operator exemplified this shift in expectations when describing their customers, which included several of the big technology companies including Amazon and Google: “They have a pace in their business that is really high and for us to keep up with their requirements and at the same time beat all our competitors we just need to be more automated”. They stated that while other customers may be more flexible and understand that instantiating a new service takes time, the “Big 5” expect services in hours rather than days and weeks.
  • Automation can enable operators to do more, such as play higher up the value chain. External partners have an expectation that telcos are highly skilled at handling data and are highly automated, particularly within the network domain. It is only through investing in internal automation efforts that operators will be able to position themselves as respected partners for services above and beyond pure connectivity. An example of success here would be the Finnish operator Elisa. They invested in automation capabilities for their own network – but subsequently have been able to monetise this externally in the form of Elisa Automate.
    A further example would be STL Partners’ vision of the Coordination Age. There is a role for telcos to play further up the value chain in coordinating across ecosystems – which will ultimately enable them to unlock new verticals and new revenue growth. The telecoms industry already connects some organisations and ecosystems together, so it’s in a strong position to play this coordinating role. But, if they wish to be trusted as ecosystem coordinators, they must first prove their pedigree in these core skills. Or, in other words, if you can’t automate your own systems, customers won’t trust you to be key partners in trying to automate theirs.
  • Automation can free up resource for service innovation. If operators are going to do more, and play a role beyond connectivity, they need to invest more in service innovation. Equally, they must also learn to innovate at a much lower cost, embracing automation alongside principles like agile development and fast fail mentalities. To invest more in service innovation, operators need to reallocate resources from other areas of their business – as most telcos are no longer rapidly growing, resource must be freed up from elsewhere.
    Reducing operating costs is a key way that operators can enable increased investment in innovation – and automation is a key way to achieve this.

A3 can drive savings to redistribute to service innovation

Source: Telecoms operator accounts, STL Partners estimates and analysis

  • 5G won’t fulfil its potential without automation. 5G standards mean that automation is built into the design from the bottom up. Most operators believe that 5G will essentially not be possible without being highly automated, particularly when considering next generation network services such as dynamic network slicing. On top of this, there will be a ranging need for automation outside of the standards – like for efficient cell-site deployment, or more sophisticated optimisation efforts for energy efficiency. Therefore, the capex investment in 5G is a major trigger to invest in automation solutions.
  • Intent-based network automation is a maturing domain. Newer technologies, like artificial intelligence and machine learning, are increasing the capabilities of automation. Traditional automation (such as robotic process automation or RPA) can be used to perform the same tasks as previously were done manually (such as inputting information for VPN provisioning) but in an automated fashion. To achieve this, rules-based scripts are used – where a human inputs exactly what it is they want the machine to do. In comparison, intent-based automation enables engineers to define a particular task (e.g. connectivity between two end-points with particular latency, bandwidth and security requirements) and software converts this request into lower level instructions for the service bearing infrastructure. You can then monitor the success of achieving the original intent.
    Use of AI and ML in conjunction with intent-based automation, can enable operators to move from automation ‘to do what humans can do but faster and more accurately’, to automation to achieve outcomes that could not be achieved in a manual way. ML and AI has a particular role to play in anomaly detection, event clustering and predictive analytics for network operations teams.
    While you can automate without AI and ML, and in fact for many telcos this is still the focus, this new technology is increasing the possibilities of what automation can achieve. 40% of our interviewees had network automation use cases that made some use of AI or ML.
  • Network virtualisation is increasing automation possibilities. As networks are increasingly virtualised, and network functions become software, operators will be afforded a greater ability than ever before to automate management, maintenance and orchestration of network services. Once networks are running on common computing hardware, making changes to the network is, in theory, purely a software change. It is easy to see how, for example, SDN will allow automation of previously human-intensive maintenance tasks. A number of operators have shared that their teams and/or organisations as a whole are thinking of virtualisation, orchestration and automation as coming hand-in-hand.

This report focuses on the opportunities and challenges in network automation. In the future, STL Partners will also look to more deeply evaluate the implications of network automation for governments and regulators, a key stakeholder within this ecosystem.

Table of Contents

  • Executive Summary
    • End-to-end network automation
    • A key opportunity: 6 reasons to focus on network automation now
    • Key recommendations for operators to drive their network automation journey
    • There are challenges operators need to overcome
    • This paper explores a range of network automation use cases
    • STL Partners: Next steps
  • Automation, analytics and AI: A3 unlocks value for operators
    • The time is ripe for network automation now
  • Looking to the future: Operators’ strategy and ambitions
    • Defining end-to-end automation
    • Defining ambitions
  • State of the industry: Network automation today
    • Which networks and what use cases: the breadth of network automation today
    • Removing the human? There is a continuum within automation use cases
    • Strategic challenges: How to effectively prioritise (network) automation efforts
    • Challenges to network automation– people and culture are key to success
  • Conclusions
    • Recommendations for vendors (and others in the ecosystem)
    • Recommendations for operators

Request STL research insights overview pack

A3 for telcos: Mapping the financial value

What is analytics, AI and automation worth to telecoms operators?

This report is the second in a two-part series mapping the process and assessing the financial value of automation, analytics and artificial intelligence (AI) in telecoms. In the first report, The value of analytics, automation and AI for telcos – Part 1: The telco A3 application map, we outlined which type of technology was best suited to which processes across a telco’s operations.

In this report, we assess the financial value of each of the operational areas, in dollar terms, for an average telco. Based on our assessment of operator financials and operational KPIs, the figure below outlines our assumptions on the characteristics of an “average” telco used as the basis for our financial modelling. The characteristics of this telco are as shown below, with a slight skew towards developed market operator characteristics since this is currently where most industry proof points used in our modelling have been implemented.

The characteristics of an average telco

characteristics of an average telco

Source: STL Partners, Charlotte Patrick Consult

The first report in the series analysed how each A3 technology could be applied similarly across different functional units of a telecoms operator, e.g. machine learning or automation each have similar processes in network management, channel management and sales and marketing.

Evaluating AI and automation use cases in four domains

To measure financial impact, this report returns to a traditional breakdown of value by functional unit within the telco, breaking down into four key areas:

  1. Network operations: Network deployment, management and maintenance, and revenue management
  2. Fraud: Including services, online, and internal fraud risks
  3. Customer care: Including all assisted and unassisted channels
  4. Marketing and sales: Understanding customers, managing products, marketing programs, lead management and sales processes.

Through an assessment of nearly 150 individual process areas across a telecoms operator’s core operations, we estimate that A3 can deliver the average telco more than $1 billion dollars in value per year, through a combination of revenue uplift and opex and capex savings, equivalent to 7% of total annual revenues.

As illustrated below, core network operations management accounts for by far the greatest proportion of the value.

The relative value of automation, AI and analytics across telco operations

The relative value of AI, automation and analytics across telco operations

Source: STL Partners, Charlotte Patrick Consult

This likely still underrepresents the total, long term potential value of A3 to telcos, since this first iteration does not model the value of A3 processes in areas less unique to telecoms, including supply chain, finance, IT and HR. No doubt that even within the core area of operations, there are potential process areas that have yet to be discovered or proven, and which we have overlooked in our initial attempt to model the value of A3 to telcos. Meanwhile, this is focused purely on telco’s internal operations so also excludes any potential revenue uplift from new A3-enabled services, such as data monetisation or development of AI-as-a-service type solutions.

That said, operators cannot implement all of these processes at once. The enormous challenge of restructuring processes to be more automated and data-centric, putting in place the data management and analytics capabilities, training employees and acquiring new skills, among many others, means that while many leading telcos are well on their way to capturing this value in some areas, very few – if any – have implemented A3 across all process areas. As a benchmark, Telefónica is an industry leader in leveraging automation and AI to improve operational efficiency, and in 2019 it reported total operational savings of €429mn across the entire group. While this is primarily focused on customer facing channels, so likely excludes the value of A3 in many network operations processes, for instance energy efficiency which is delivering significant value to Telefónica and others, it suggests there remains lots of value still to capture.

Request a report extract

Methodology

The financial modelling for the value of A3 was done through an individual assessment of each of the 150+ process areas to understand the main activities within that area of operations, and how automation, analytics and/or machine learning and other AI technologies could be used within those activities. From there, we assess the value of integrating these technologies to existing operational functions to make them more efficient and effective. This means that we do not attribute any additional value to telcos from implementing new technologies that include A3 as a core element of their functionality, e.g. a multi-domain service orchestrator, implemented as part of software-defined networking.

Our bottom up assessment of each process is also validated through real-world proof points from operators or vendors. This means that more speculative areas of A3 application in operators are calculated to offer relatively limited value. As more proof points emerge, we will incorporate them into future iterations.

Table of contents

  • Executive Summary
    • Where is the largest financial benefit from A3?
    • What should telcos prioritise in the short term?
    • How long will it take for telcos to realise this value?
    • What next?
  • Introduction
    • Methodology
  • Breaking down the value of A3 by operational area
    • Network, OSS and BSS
    • Fraud management
    • Care and commercial channels
    • Marketing and sales
  • Conclusions and recommendations

Request STL research insights overview pack

Coordinating the care of the elderly

Are telcos ready to enable digital health?

The world has been talking about connected healthcare – the use of in-home and wearable systems to monitor people’s condition – for a long time. Although adoption to date has been piecemeal and limited, the rapid rise in the number of elderly people is fuelling demand for in-home and wearable monitoring systems. The rapid spread of the Covid-19 virus is putting the world’s healthcare systems under huge strain, further underlining the need to reform the way in which many medical conditions are diagnosed and treated.

This report explores whether telcos now have the appetite and the tools they need to serve this very challenging, but potentially rewarding market. With the advent of the Coordination Age (see STL Partners report: Telco 2030: New purpose, strategy and business models for the Coordination Age), telcos could play a pivotal role in enabling the world’s healthcare systems to become more sustainable and effective.

This report considers demographic trends, the forces changing healthcare and the case for greater use of digital technologies to monitor chronic conditions and elderly people. It explores various implementation options and some of the healthcare-related activities of Tele2, Vodafone, Telefónica and AT&T, before drawing conclusions and recommending some high-level actions for telcos looking to support healthcare for the elderly.

This executive briefing builds on previous STL Partners reports including:

Request a report extract

Why healthcare needs to change

During the twentieth century, life expectancy in most countries in the world rose dramatically.  This was down to advances in medical science and diagnostic technology, as well as rising awareness about personal and environmental hygiene, health, nutrition, and education. Average global life expectancy continues to rise, increasing from 65.3 years in 1990 to 71.5 years in 2013.  In some countries, the increase in lifespans has been dramatic. The life expectancy for a Chilean female has risen to 82 years today from 33 years in 1910, according to the World Health Organization (WHO).

Figure 1: Across the world, average life expectancy is rising towards 80

raising lift expectancy to 2050

Source: The UN

Clearly, the increase in the average lifespan is a good thing. But longer life expectancy, together with falling birth rates, means the population overall is aging rapidly, posing a major challenge for the world’s healthcare systems. According to the WHO, the proportion of the world’s population over 60 years old will double from about 11% to 22% between 2000 and 2050, equivalent to a rise in the absolute number of people over 60 from 605 million to an extraordinary two billion. Between 2012 and 2050, the number of people over 80 will almost quadruple to 395 million, according to the WHO. That represents a huge increase in the number of elderly people, many of whom will require frequent care and medical attention. For both policymakers and the healthcare industry, this demographic time bomb represents a huge challenge.

Rising demand for continuous healthcare

Of particular concern is the number of people that need continuous healthcare. About 15% of the world’s population suffers from various disabilities, with between 110 million and 190 million adults having significant functional difficulties, according to the WHO. With limited mobility and independence, it can be hard for these people to get the healthcare they need.

As the population ages, this number will rise and rise. For example, the number of Americans living with Alzheimer’s disease, which results in memory loss and other symptoms of dementia, is set to rise to 16 million by 2050 from five million today, according to the Alzheimer’s Association.

The growth in the number of older people, combined with an increase in sedentary lifestyles and diets high in sugars and fats, also means many more people are now living with heart disease, obesity, diabetes and asthma. Furthermore, poor air quality in many industrial and big cities is giving rise to cancer, cardiovascular and respiratory diseases such as asthma, and lung diseases. Around 235 million people are currently suffering from asthma and about 383,000 people died from asthma in 2015, according to the WHO.

Half of all American adults have at least one chronic condition with one in three adults suffering from multiple chronic conditions, according to the National Institutes of Health (NIH). Most other rich countries are experiencing similar trends, while middle-income countries are heading in the same direction. In cases where a patient requires medical interventions, they may have to travel to a hospital and occupy a bed, at great expense. With the growing prevalence of chronic conditions, a rising proportion of GDP is being devoted to healthcare. Only low-income countries are bucking this trend (see Figure 2).

Figure 2: Spending on healthcare is rising except in low income countries

Public health as % of government spending WHO

Public health spending as % of GDP WHO

Source: The WHO

However, there is a huge difference in absolute spending levels between high-income countries and the rest of the world (see Figure 3). High-income countries, such as the U.S., spend almost ten times as much per capita as upper middle-income countries, such as Brazil. At first glance, this suggests the potential healthcare market for telcos is going to be much bigger in Europe, North America and developed Asia, than for telcos in Latin America, developing Asia and sub-Saharan Africa. Yet these emerging economies could leapfrog their developed counterparts to adopt connected self-managed healthcare systems, as the only affordable alternative.

Figure 3: Absolute health spending in high income countries is far ahead of the rest

per capita health spending by country income levelSource: The WHO

The cost associated with healthcare services continues to rise due to the increasing prices of prescription drugs, diagnostic tools and in-clinic care. According to the U.S. Centers for Disease Control and Prevention, 90% of the nation’s US$3.3 trillion annual healthcare expenditure is spent on individuals with chronic and mental health conditions.

On top of that figure, the management of chronic conditions consumes an enormous amount of informal resources. As formal paid care services are expensive, many older people rely on the support of family, friends or volunteers calling at their homes to check on them and help them with tasks, such as laundry and shopping. In short, the societal cost of managing chronic conditions is enormous.

The particular needs of the elderly

Despite the time and money being spent on healthcare, people with chronic and age-related conditions can be vulnerable. While most elderly people want to live in their own home, there are significant risks attached to this decision, particularly if they live alone. The biggest danger is a fall, which can lead to fractures and, sometimes, lethal medical complications. In the U.S., more than one in four older people fall each year due to illness or loss of balance, according to the U.S. Centers for Disease Control and Prevention. But less than half tell their doctor. One out of five falls causes a serious injury, such as broken bones or a head injury. In 2015, the total medical costs for falls was more than US$50 billion in the U.S. Beyond falls, another key risk is that older people neglect their own health. A 2016 survey of 1,000 U.K. consumers by IT solutions company Plextek, found that 42% of 35- to 44-year-olds are concerned that their relatives aren’t telling them they feel ill.

Such concerns are driving demand for in-home and wearable systems that can monitor people in real-time and then relay real-time location and mobility information to relatives or carers. If they are perceived to be reliable and comprehensive, such systems can provide peace of mind, making home-based care a more palatable alternative for both patients and their families.

Request STL's digital health insights pack

Table of contents

  • Executive Summary
    • Barriers to more in-home healthcare
  • Introduction
  • Why healthcare needs to change
    • Rising demand for continuous healthcare
    • The particular needs of the elderly
    • Shift to value-based care
    • Demands for personalised healthcare and convenience
  • How healthcare is changing
    • Barriers to more in-home healthcare
  • Implementation options
    • Working with wearables
    • Cameras and motion sensors
    • The connectivity
    • Analysing the data
  • How telcos are tackling healthcare
    • KPN: Covering most of the bases
    • Tele2 and Cuviva: Working through healthcare centres
    • Vodafone and Vision: An expensive system for Alzheimer’s
    • Telefónica’s Health Moonshot
    • AT&T: Leveraging a long-standing brand
  • Conclusions and recommendations
    • Recommendations

The Industrial IoT: What’s the telco opportunity?

The Industrial IoT is a confusing world

This report is the final report in a mini-series about the Internet for Things (I4T), which we see as the next stage of evolution from today’s IoT.

The first report, The IoT is dead: Long live the Internet for Things, outlines why today’s IoT infrastructure is insufficient for meeting businesses’ needs. The main problem with today’s IoT is that every company’s data is locked in its own silo, and one company’s solutions are likely deployed on a different platform than their partners’. So companies can optimise their internal operations, but have limited scope to use IoT to optimise operations involving multiple organisations.

The second report, Digital twins: A catalyst of disruption in the Coordination Age, provides an overview of what a digital twin is, and how they can play a role in overcoming the limitations of today’s IoT industry.

This report looks more closely at the state of development of enterprise and industrial IoT and the leading players in today’s IoT industry, which we believe is a crucial driver of the Coordination Age. In the Coordination Age, we believe the crucial socio-economic need in the world – and therefore the biggest business opportunity – is to make better use of our resources, whether that is time, money, or raw materials. Given the number of people employed in and resources going through industrial processes, figuring out what’s needed to make the industrial IoT reach its full potential is a big part of making this possible.

Three types of IoT

There are three ways of dividing up the types of IoT applications. As described by IoT expert Stacey Higginbotham, each group has distinct needs and priorities based on their main purpose:

  1. Consumer IoT: A connected device, with an interactive app, that provides an additional service to the end user compared with an unconnected version of the device. The additional service is enabled by the insights and data gathered from the device. The key priority for consumer devices is low price point and ease of installation, given most users’ lack of technical expertise.
  2. Enterprise IoT: This includes all the devices and sensors that enterprises are connecting to the internet, e.g. enterprise mobility and fleet tracking. Since every device connected to an enterprise network is a potential point of vulnerability, the primary concern of enterprise IoT is security and device management. This is achieved through documentation of devices on enterprise networks, prioritisation of devices and traffic across multiple types of networks, e.g. depending on speed and security requirements, and access rights controls, to track who is sharing data with whom and when.
  3. Industrial IoT: This field is born out of industrial protocols such as SCADA, which do not currently connect to the internet but rather to an internal control and monitoring system for manufacturing equipment. More recently, enterprises have enhanced these systems with a host of devices connected to IP networks through Wi-Fi or other technologies, and linked legacy monitoring systems to gateways that feed operational data into more user-friendly, cloud-based monitoring and analytics solutions. At this point, the lines between Industrial IoT and Enterprise IoT blur. When the cloud-based systems have the ability to control connected equipment, for instance through firmware updates, security to prevent malicious or unintended risks is paramount. The primary goals in IIoT remain to control and monitor, in order to improve operational efficiency and safety, although with rising security needs.

The Internet for Things (I4T) is in large part about bridging the divide between Enterprise and Industrial IoT. The idea is to be able to share highly sensitive industrial information, such as a change in operational status that will affect a supply chain, or a fault in public infrastructure like roads, rail or electricity grid, that will affect surroundings and require repairs. This requires new solutions that can coordinate and track the movement of Industrial IoT data into Enterprise IoT insights and actions.

Understandably, enterprises are way of opening any vulnerabilities into their operations through deeper or broader connections, so finding a secure way to bring about the I4T is the primary concern.

The proliferation of IoT platforms

Almost every major player in the ICT world is pitching for a role in both Enterprise and Industrial IoT. Most largescale manufacturers and telecoms operators are also trying to carve out a role in the IoT industry.

By and large, these players have developed specific IoT solutions linked to their core businesses, and then expanded by developing some kind of “IoT platform” that brings together a broader range of capabilities across the IT stack necessary to provide end-to-end IoT solutions.
The result is a hugely complex industry with many overlapping and competing “platforms”. Because they all do something different, the term “platform” is often unhelpful in understanding what a company provides.

A company’s “IoT platform” might comprise of any combination of these four layers of the IoT stack, all of which are key components of an end-to-end solution:

  1. Hardware: This is the IoT device or sensor that is used to collect and transmit data. Larger devices may also have inbuilt compute power enabling them to run local analysis on the data collected, in order to curate which data need to be sent to a central repository or other devices.
  2. Connectivity: This is the means by which data is transmitted, including location-based connectivity (Bluetooth, Wi-Fi), to low power wide area over unlicensed spectrum (Sigfox, LoRa), and cellular (NB-IoT, LTE-M, LTE).
  3. IoT service enablement: This is the most nebulous category, because it includes anything that sits as middleware in between connectivity and the end application. The main types of enabling functions are:
    • Cloud compute capacity for storing and analysing data
    • Data management: aggregating, structuring and standardising data from multiple different sources. There are sub-categories within this geared towards specific end applications, such as product or service lifecycle management tools.
    • Device management: device onboarding, monitoring, software updates, and security. Software and security management are often broken out as separate enablement solutions.
    • Connectivity management: orchestrating IoT devices over a variety of networks
    • Data / device visualisation: This includes graphical interfaces for presenting complex data sets and insights, and 3D modelling tools for industrial equipment.
  4. Applications: These leverage tools in the IoT enablement layer to deliver specific insights or trigger actions that deliver a specific outcome to end users, such as predictive maintenance or fleet management. Applications are usually tailored to the specific needs of end users and rarely scale well across multiple industries.

Most “IoT platforms” combine at least two layers across this IoT stack

graphic of 4 layers on the IoT stack

Source: STL Partners

There are two key reasons why platforms offering end-to-end services have dominated the early development of the IoT industry:

  • Enterprises’ most immediate needs have been to have greater visibility into their own operations and make them more efficient. This means IoT initiatives have been driven primarily by business owners, rather than technology teams, who often don’t have the skills to piece together multiple different components by themselves.
  • Although the IoT as a whole is a big business, each individual component to bringing a solution together is relatively small. So companies providing IoT solutions – including telcos – have attempted to capture a larger share of the value chain in order to make it a better business.

Making sense of the confusion

It is a daunting task to work out how to bring IoT into play in any organisation. It requires a thorough re-think of how a business operates, for a start, then tinkering with (or transforming) its core systems and processes, depending on how you approach it.

That’s tricky enough even without the burgeoning market of self-proclaimed “leaders of industrial IoT” and technology players’ “IoT platforms”.

This report does not attempt to answer “what is the best way / platform” for different IoT implementations. There are many other resources available that attempt to offer comparisons to help guide users through the task of picking the right tools for the job.

The objective here is to gain a sense of what is real today, and where the opportunities and gaps are, in order to help telecoms operators and their partners understand how they can help enterprises move beyond the IoT, into the I4T.

 

Table of contents

  • Executive Summary
  • Introduction
    • Three types of IoT
    • The proliferation of IoT platforms
    • Making sense of the confusion
  • The state of the IoT industry
    • In the beginning, there was SCADA
    • Then there were specialised industrial automation systems
    • IoT providers are learning about evolving customer needs
  • Overview of IoT solution providers
    • Generalist scaled IT players
    • The Internet players (Amazon, Google and Microsoft)
    • Large-scale manufacturers
    • Transformation / IoT specialists
    • Big telco vendors
    • Telecoms operators
    • Other connectivity-led players
  • Conclusions and recommendations
    • A buyers’ eye view: Too much choice, not enough agility
    • How telcos can help – and succeed over the long term in IoT

New age, new control points?

Why control points matter

This executive briefing explores the evolution of control points – products, services or roles that give a company disproportionate power within a particular digital value chain. Historically, such control points have included Microsoft’s Windows operating system and Intel’s processor architecture for personal computers (PCs), Google’s search engine and Apple’s iPhone. In each case, these control points have been a reliable source of revenues and a springboard into other lucrative new markets, such as productivity software (Microsoft) server chips (Intel), display advertising (Google) and app retailing (Apple).

Although technical and regulatory constraints mean that most telcos are unlikely to be able to build out their own control points, there are exceptions, such as the central role of Safaricom’s M-Pesa service in Kenya’s digital economy. In any case, a thorough understanding of where new control points are emerging will help telcos identify what their customers most value in the digital ecosystem. Moreover, if they move early enough to encourage competition and/or appropriate regulatory intervention, telcos could prevent themselves, their partners and their customers from becoming too dependent on particular companies.

The emergence of Microsoft’s operating system as the dominant platform in the PC market left many of its “partners” struggling to eke out a profit from the sale of computer hardware. Looking forward, there is a similar risk that a company that creates a dominant artificial intelligence platform could leave other players in various digital value chains, including telcos, at their beck and call.

This report explores how control points are evolving beyond simple components, such as a piece of software or a microprocessor, to become elaborate vertically-integrated stacks of hardware, software and services that work towards a specific goal, such as developing the best self-driving car on the planet or the most accurate image recognition system in the cloud. It then outlines what telcos and their partners can do to help maintain a balance of power in the Coordination Age, where, crucially, no one really wants to be at the mercy of a “master coordinator”.

The report focuses primarily on the consumer market, but the arguments it makes are also applicable in the enterprise space, where machine learning is being applied to optimise specialist solutions, such as production lines, industrial processes and drug development. In each case, there is a danger that a single company will build an unassailable position in a specific niche, ultimately eliminating the competition on which effective capitalism depends.

Enter your details below to request an extract of the report


Control points evolve and shift

A control point can be defined as a product, service or solution on which every other player in a value chain is heavily dependent. Their reliance on this component means the other players in the value chain generally have to accept the terms and conditions imposed by the entity that owns the control point. A good contemporary example is Apple’s App Store – owners of Apple’s devices depend on the App Store to get access to software they need/want, while app developers depend on the App Store to distribute their software to the 1.4 billion Apple devices in active use. This pivotal position allows Apple to levy a controversial commission of 30% on software and digital content sold through the App Store.

But few control points last forever: the App Store will only continue to be a control point if consumers continue to download a wide range of apps, rather than interacting with online services through a web browser or another software platform, such as a messaging app. Recent history shows that as technology evolves, control points can be sidestepped or marginalised. For example, Microsoft’s Windows operating system and Internet Explorer browser were once regarded as key control points in the personal computing ecosystem, but neither piece of software is still at the heart of most consumers’ online experience.

Similarly, the gateway role of Apple’s App Store looks set to be eroded over time. Towards the end of 2018, Netflix — the App Store’s top grossing app — no longer allowed new customers to sign up and subscribe to the streaming service within the Netflix app for iOS across all global markets, according to a report by TechCrunch. That move is designed to cut out the expensive intermediary — Apple. Citing data compiled by Sensor Tower, the report said Netflix would have paid Apple US$256 million of the US$853 million grossed by its 2018 the Netflix iOS app, assuming a 30% commission for Apple (however, after the first year, Apple’s cut on subscription renewals is lowered to 15%).

TechCrunch noted that Netflix is following in the footsteps of Amazon, which has historically restricted movie and TV rentals and purchases to its own website or other “compatible” apps, instead of allowing them to take place through its Prime Video app for iOS or Android. In so doing, Amazon is preventing Apple or Google from taking a slice of its content revenues. Amazon takes the same approach with Kindle e-books, which also aren’t offered in the Kindle mobile app. Spotify has also discontinued the option to pay for its Premium service using Apple’s in-app payment system.

Skating ahead of the puck

As control points evolve and shift, some of today’s Internet giants, notably Alphabet, Amazon and Facebook, are skating where the puck is heading, acquiring the new players that might disrupt their existing control points. In fact, the willingness of today’s Internet platforms to spend big money on small companies suggests they are much more alert to this dynamic than their predecessors were. Facebook’s US$19 billion acquisition of messaging app WhatsApp, which has generated very little in the way of revenues, is perhaps the best example of the perceived value of strategic control points – consumers’ time and attention appears to be gradually shifting from traditional social into messaging apps, such as WhatsApp, or hybrid-services, such as Instagram, which Facebook also acquired.

In fact, the financial and regulatory leeway Alphabet, Amazon, Facebook and Apple enjoy (granted by long-sighted investors) almost constitutes another control point. Whereas deals by telcos and media companies tend to come under much tougher scrutiny and be restricted by rigorous financial modelling, the Internet giants are generally trusted to buy whoever they like.

The decision by Alphabet, the owner of Google, to establish its “Other Bets” division is another example of how today’s tech giants have learnt from the complacency of their predecessors. Whereas Microsoft failed to anticipate the rise of tablets and smart TVs, weakening its grip on the consumer computing market, Google has zealously explored the potential of new computing platforms, such as connected glasses, self-driving cars and smart speakers.

In essence, the current generation of tech leaders have taken Intel founder Andy Grove’s famous “only the paranoid survive” mantra to heart. Having swept away the old order, they realise their companies could also easily be side-lined by new players with new ways of doing things. Underlining this point, Larry Page, founder of Google, wrote in 2014: “Many companies get comfortable doing what they have always done, making only incremental changes. This incrementalism leads to irrelevance over time, especially in technology, where change tends to be revolutionary, not evolutionary. People thought we were crazy when we acquired YouTube and Android and when we launched Chrome, but those efforts have matured into major platforms for digital video and mobile devices and a safer, popular browser.”

Table of contents

  • Executive Summary
  • Introduction
  • What constitutes a control point?
    • Control points evolve and shift
    • New kinds of control points
  • The big data dividend
    • Can incumbents’ big data advantage be overcome?
    • Data has drawbacks – dangers of distraction
    • How does machine learning change the data game?
  • The power of network effects
    • The importance of the ecosystem
    • Cloud computing capacity and capabilities
    • Digital identity and digital payments
  • The value of vertical integration
    • The machine learning super cycle
    • The machine learning cycle in action – image recognition
  • Tesla’s journey towards self-driving vehicles
    • Custom-made computing architecture
    • Training the self-driving software
    • But does Tesla have a sustainable advantage?
  • Regulatory checks and balances
  • Conclusions and recommendations

Enter your details below to request an extract of the report


Telco AI: How to organise and partner for maximum success

Not a passing fad: AI is becoming a core capability for telcos

Artificial intelligence (AI) has become a key enabler of the digital transformation journey for service providers in the telecoms industry, providing them with the insights and capabilities they need to be more agile and take a more software-centric approach to their role.

The document was researched and written independently by STL Partners, supported by Nokia. STL’s conclusions are entirely independent and build on ongoing research into the future of telecoms. STL Partners has been writing about telcos’ AI opportunities since 2016, looking first at how AI might improve the customer experience and then at the critical role AI might play in the future of network operations.

In this report, we provide a comprehensive overview of the state of AI in the telecoms industry. Supported by nearly a dozen in-depth interviews plus an online survey of more than 50 leading telcos around the world, we explore where the industry is looking to progress and how it is planning to do so — and identify the strategic and business opportunities that are being enabled by AI.

This report will be followed by a sequel that quantifies some of the business outcomes telcos can expect from specific AI application areas. In the coming months, we will also publish a report discussing how AI technology is evolving and presenting our vision of the telco AI roadmap.

What is artificial intelligence?

Before going any further, it is important to clarify what we mean by “artificial intelligence”. To us, AI is about using computing capabilities to perform tasks traditionally associated with humans (such as inference, planning, anticipation, prediction and learning) in human-like ways (e.g., autonomous, adaptive). Our definition incorporates machine learning (ML), which we define as a subset of AI that focuses on the ability of machines to receive datasets and adapt responses in pursuit of a goal.

These definitions attempt to encapsulate the distinction between AI and other forms of rules-based automation — although we acknowledge that in practice these lines are easily blurred.

Practically speaking, AI sits on a continuum of other related technologies and concepts, which we have covered at length in our previous reports. Figure 1 illustrates this continuum and depicts the stages we expect telcos will have to go through as they to move from manual to automated and then to AI-augmented processes.

Figure 1: Moving toward AI

The progression of AI maturity in four steps

Source: STL Partners

A long-term ambition for many telcos is to reach the orange zone in Figure 1: a state in which their systems and processes run and learn from themselves with human input limited to the setting of desired business goals. Beyond the targeted use of ML in certain applications, however, the industry and society as a whole are far from realising that ambition. It is still unclear what fully autonomous systems in a telco might look like in practice, let alone whether they will ever be achievable.

Today, most telcos are still figuring out how to play in the blue zone. They’re using targeted data analysis to inform largely human-led decision-making processes, or they’ve implemented some fixed-policy automation where machines follow a script written and inputted by a human. This is valuable work, but it is not the focus of this report. Instead, we focus on the middle section of Figure 1: on those fledging opportunities that move beyond rules-based automation and into the realm of ML-supported automation

Cutting through the hype

AI has generated considerable industry noise and media attention — so much so, in fact, that a recent survey of leading telcos awarded AI the title of “most overhyped emerging technology”. We believe this hype originates in a general lack of understanding of what AI is (and is not), as well as unrealistic expectations about what it can do for a business, how quickly it can be deployed, and how much ongoing work will be needed to manage it. While there is consensus that the technology has great potential, many telcos doubt it will deliver everything that has been promised up to now.

For those disillusioned by the hype, it is worth noting the impact of AI is much likelier to be evolutionary than revolutionary. The line between automation and AI is blurred; so, too, is the progression between the two. While AI has the potential to unlock new business opportunities, realising that potential will require patience and long-term investment.

And yet, the truth is that telcos are uniquely positioned to take full advantage of AI technology — largely because they’re already used to dealing with the huge volumes of data AI relies on. When telcos automate systems, networks and processes — particularly with the injection of AI — they benefit from feedback loops that further improve those automated processes. This drives simplicity in an industry rife with complexity.

The digital transformation we all talk about depends on driving out complexity and becoming more agile, and the only way to do that is by automating intelligently. Looking ahead to the launch of 5G, it will become impossible for telcos to manage billions of connected devices without AI assistance.

Telcos’ current AI focus: Improving speed and efficiency

Key learnings on telco AI initiatives

Through our research, we have identified five primary domains of activity for telcos looking to make use of AI. The first three broadly relate to business process improvement, with the end goal of reducing costs and improving efficiency.

  1. Optimising existing networks and operations. Telcos are using AI not only for network planning and optimisation, but also to improve their human resources, accounting and fraud-management functions. For example, Telefónica has built an ML model capable of monitoring the status of the network, predicting possible failures and an optimising maintenance routes.[1] This has been particularly important in its rollout and maintenance of networks across rural Latin America, where it can take an engineer up to a day to travel to the site of a network fault.
  2. Improving sales and marketing activity. This includes upselling, cross-selling and agent augmentation. Globe Telecom, for example, has created a data-management platform that collates network signal information alongside information from billing and payment systems to provide personalised offers to its mobile customers.[2]
  3. Improving the customer experience. This includes use cases such as fault resolution, churn management, chatbots and virtual assistants. Vodafone has developed the chatbot TOBi, for example, which can handle 70 percent of customer requests and employs ML technology to further improve the support it offers to customers.[3]

The remaining two domains focus on using AI to enable new ways of working that go beyond a telco’s core connectivity offering, with a focus on growing revenues.

  1. Driving (and monetising) customer data. AI can help telcos aggregate massive volumes of anonymised customer data that can then be sold to third parties. Singtel’s DataSpark has taken a step down this data-as-a-service route, providing access to GPS and mobile network data that other organisations can incorporate into their applications and services.[4]
  2. Enabling or supporting new services. This includes cybersecurity and predictive analytics. As an example, AT&T is using ML to quickly identify normal and abnormal activity in it networks.[5] This sort of solution could be sold as a managed service to other enterprises in the future, unlocking a new revenue stream.

Contents of the full report include:

  • Executive Summary
  • Not a passing fad: AI is becoming a core capability for telcos
  • What is artificial intelligence?
  • Cutting through the hype 8
  • Telcos’ current AI focus: Speed and efficiency
  • How are telcos using AI today?
  • Sharing is caring: How telco AI initiatives are organised
  • Centralised AI initiatives
  • Cross-functional R&D units
  • Individual AI initiatives
  • The stumbling blocks for AI implementation — and how to get around them
  • AI initiatives need to be powered by high-quality data
  • Data governance is an essential requirement
  • Exploring the link between data maturity and AI success
  • The tricky transition from the lab to in-field deployment
  • Accept failure and embrace innovation
  • Revamp partnership strategies
  • New challenges, new expectations
  • Finding the impact: How telcos assess the benefits of AI
  • Different types of telcos, different levels of AI maturity
  • Conclusion

Figures:

  1. Moving toward AI
  2. Telco AI initiatives by domain
  3. Centrally coordinated AI initiatives are more likely to scale
  4. Poor data and a lack of internal skills are key challenges
  5. Telcos struggle with data management at every step of the AI journey
  6. Issues with data governance do not preclude AI implementation
  7. Only 1 in 5 AI projects has advanced to live deployment
  8. Collaborative partnering is key to AI success
  9. Nearly half of telcos have not gone live with AI
  10. Fixed-line and wholesale operators lag behind


[1] Source: Telefónica

[2] Source: Cloudera

[3] Source: Vodafone

[4] Source: DataSpark

[5] Source: AT&T

Network AI: The state of the art

Introduction

This report is part of a series exploring how telecoms operators can leverage artificial intelligence (AI) to improve their business operations, from customer experience to new services. Previous reports on AI in telecoms include:

This report explores the applications of AI for network operations, detailing the prerequisites and stages to implementing AI and automation in networks, real-world examples of what some telcos have done already, and their potential value across different application areas.

We divide the applications for AI in telecoms networks into three main categories:

  • Fault detection, prediction and resolution: speeding up the process of identifying and resolving network faults, including predictive maintenance. This also includes identifying and mitigating network security risks, although security is a highly specialised field that merits its own report, so we do not cover it in detail here.
  • Network optimisation: optimising the use of network resources to mitigate the impact of network faults and adapt to or anticipate changes in demand. This is also the foundation for automated service provisioning in software defined networks, while insights on network usage and traffic could be valuable for new service creation.
  • Network planning and upgrades: optimising new infrastructure planning as well as the transition from legacy to next generation network solutions.

The first area is critical for all telcos, since service impairments are an inevitable element of running a network. The second is of immediate value for telcos that are still in the process of expanding existing network coverage and density, since it can enable operators to use their existing resources more efficiently. However, it is also increasingly tied into the first area of fault detection, since a large part of the fault resolution process is finding ways to re-route traffic from underperforming to underused assets, a process that is made easier with the adoption of SDN and NFV – processes can only be automated if they are software-based.

Compared with the first two categories, using AI for smarter network planning and upgrades is a nascent field. This is partially because many Tier 1 operators, who are leading the charge in adoption of AI elsewhere in network and business operations, completed the bulk of 4G deployments and have not yet fully embarked on 5G deployments. However, this report also looks at some innovative applications of image recognition models for network expansion in emerging markets.

While most of the data used for training and informing AI systems across network operations comes from operators’ own networks, telcos are also beginning to tap into new data sources to further refine their decision-making, such as using drones and image recognition to inspect towers, weather patterns and social media data.

Laying the foundations for AI in telecoms networks

Before jumping into how telcos are implementing AI for fault detection and resolution and in network operations, it is important to clarify what we mean by AI, and lay out the pre-requisites for any meaningful use of the technology.

What counts as AI? From automation to advanced AI

The term AI is nebulous – everyone has a different definition for it. Is it when a computer can make a faster, more accurate decision than a human?  Is it when a process is fully automated? Is it when the computer learns and continuously improves its decisions in real-time?

Wherever people draw the line between manual processes, (big) data analytics, automation and machine learning (ML) / AI, no company goes directly from manual to AI in one go. The transition is gradual. In this report we therefore use a broad definition of AI in this report, as outlined in Figure 1.

Figure 1: Not all AI is equal

Rules-based automation to machine learning

Source: STL Partners

Two transitions are happening in parallel as operators move from left to right on Figure 1. First, there is a shift towards increasingly intelligent analytics techniques, from rules-based automation, where policies outline if-then sequences of actions for the computer, to machine learning supported automation, where models are trained to fulfil an intent (a goal) based on guidelines from experts and historical data.

The second transition that occurs in the move towards more sophisticated AI systems relates to decision-making. In rules-based automation, computers don’t have any decision-making power, they can only take pre-defined actions in specific circumstances. Making the transition from telling computers how to do something to what you want them to do means giving computers decision-making power. Telcos can do this gradually, by requiring humans to verify and approve recommended decisions before they are implemented. But in the promised future 5G and ‘sliceable’ networks, human approval for routine decisions would require more network engineers than operators could profitably employ, or drastically slow down network operations. This is not just a technical issue for telcos but also a cultural one that demands clarity from management teams on the evolving role of network engineers.

Contents:

  • Executive Summary
  • Making the shift from manual operations to autonomous, intelligent networks
  • Recommendations
  • Introduction
  • Laying the foundations for AI in telecoms networks
  • What counts as AI? From automation to advanced AI
  • AI works at two levels for network operations
  • Data: The bridge between rules-based automation and ML
  • Fault detection, prediction and resolution
  • What is it worth?
  • How does it work?
  • Real-world example of a recommendation model: AT&T Tower Outage and Network Analyzer
  • Next step: From fixed to self-learning policies
  • Optimising network capacity
  • What are self-optimising networks worth?
  • Use case overview
  • How to do it
  • From self-optimising to knowledge-defined networks
  • AI for network planning
  • Telefónica case study
  • Driving automation internally versus partnering with vendors
  • Reasons for developing solutions internally
  • Reasons for partnering with a vendor
  • Vendor profiles
  • How AI fits with SDN/NFV
  • Conclusions and recommendations

Figures:

  • Figure 1: Not all AI is equal
  • Figure 2: Rules-based automation versus machine learning
  • Figure 3: A snapshot of rules-based automation versus machine learning
  • Figure 4: Overview of automation and AI in network operations
  • Figure 5: Telemetry is faster and uses less compute power than SNMP
  • Figure 6: Elisa growth of automated trouble ticket handling
  • Figure 7: Tupl results for automatic customer complaints resolution AI platform
  • Figure 8: Implementing fixed policies for fault detection and resolution
  • Figure 9: Visualisation of network alert clustering tool
  • Figure 10: A self-healing network
  • Figure 11: Elisa self-optimising network results
  • Figure 12: Elisa maintained flat capex intensity throughout 4G deployment
  • Figure 13: Finland 4G network performance, August 2018
  • Figure 14: Self-organising network example use cases
  • Figure 15: Numerous applications of machine learning and AI for 5G networks
  • Figure 16: Break self-optimising networks down into mini loops
  • Figure 17: The knowledge-defined network
  • Figure 18: Facebook TCO savings over traditional multilayer planning
  • Figure 19: Telefónica image recognition for network planning
  • Figure 20: Ciena Blue Planet overview
  • Figure 21: Google SDN layers
  • Figure 22: Overview of cross-industry initiatives relating to network AI and automation
  • Figure 23: Telefónica network automation roadmap
  • Figure 24: Overview of SK Telecom Advanced Next Generation OSS (TANGO)

AI in customer services: It’s not all about chatbots

Introduction

Internet companies like Google, Apple and Amazon and second tier digital disruptors like Airbnb, Spotify and Netflix have been refining advanced machine learning-enabled analytics to offer increasingly personalised and predictive services. This means consumers increasingly expect a seamless, personalised experience, where service providers anticipate their needs, rather than react to problems after the fact.

If telcos want to regain credibility with consumers, they must develop more personalised and frictionless customer experiences. Many telcos believe that, as for digital native companies, artificial intelligence (AI) technologies can play a crucial role in achieving their goal to reduce costs while improving services.

In this report, we assess the potential value and feasibility of different types of AI applications in customer experience and customer care and outline how telecoms operators should prioritise their efforts. It is based on primary research at STL events, and conversations with telcos, vendors and other industry players. It also builds on previous reports on big data and AI:

Defining AI: An evolution in analytics

In this and following reports, we are using AI as an all-encompassing term for advanced predictive analytics, based on machine learning technologies. Machine learning should be seen as a stage in the evolution of analytics, from basic data analysis, to big data analytics, to fully automated systems where algorithms continually learn from new data to deliver more efficient and effective responses and automated actions.

In our recent report, Big data analytics: Time to up the ante, we emphasized that as telcos have built up their big data programmes, they have gradually replaced siloed legacy infrastructure, where each department has their own databases and data sources, with a horizontally-integrated infrastructure across all departments. This shift enables telcos to run advanced analytics across much broader data sets and use machine learning algorithms to uncover new patterns and correlations across previously segregated data, often described as discovering ‘unknown unknowns’.

Within the field of machine learning, there are three types of algorithms:

  1. Supervised learning: an algorithm is trained on a large set of labelled data (e.g. photos of cats, or examples of spam mail). The algorithm learns to find patterns in the training data, which it then applies to new data samples to predict the correct answer (e.g. one email is spam, another is not). This is good for situations where historical data can predict future events, such as when a customer is likely to churn, or when a location is likely to see a peak in demand for connectivity.
  2. Unsupervised learning: an algorithm is applied to data without labels and with no specific goal, other than to find patterns or structure in the data. This is good for finding new ways to segment customers or detecting outliers, for example fraudulent activity, or an operator’s most and least profitable customers.
  3. Reinforcement learning: an algorithm is given a predefined goal and a set of allowed actions, and is then left to find the most effective way to achieve its goal through trial and error. This is most famously used in gaming, for example by Google’s AlphaGo Zero. Testing different marketing campaigns to achieve the best result is an example of a business application of reinforcement learning.

Deep learning (DL) is an extension of machine learning where there are many more layers in a neural network, allowing the system to work with much larger and more complex data sets, such as images, video, text and audio, in order to identify more subtle patterns.

In most application fields, ML and DL algorithms are not yet at the stage where they can teach themselves what to do, without any guidance and oversight from humans. Most algorithms are also trained for a specific purpose, so their applications are not easily transferrable. But even with these constraints, the benefits of AI for businesses are clear:

  • Self-improving: ML algorithms continuously learn from experience, either from inbuilt rewards systems or guidance from humans
  • High performing: AI can handle huge amounts of data and never sleeps
  • Human-like: it can interact with types of data previously indecipherable by software, enabling it to automate previously uniquely human tasks, e.g. natural language understanding, facial recognition.
  • Wide reaching: AI is suited to automation of both soft (e.g. chatbots, system control) and manual tasks (i.e. robotics)

Contents:

  • Executive Summary
  • Chatbots aren’t an easy win
  • Predictive care: an easier entry point, but with limits
  • So is AI in customer care worth it?
  • Introduction
  • Defining AI: An evolution in analytics
  • STL Partners’ AI Framework: Sizing the opportunity
  • Chatbots: Is it worth the work?
  • What is a chatbot?
  • Most chatbots are not ready to handle customers yet, and vice versa
  • Telenor: Taking an active role in the AI technology revolution
  • T-Mobile Austria (Deutsche Telekom): A more personalised customer experience
  • Recommendations
  • Predictive care & agent assist
  • AT&T is prioritising predictive care
  • Lots of room for improvement in handling customer calls
  • But predictive care can’t solve everything
  • Conclusions

Figures:

  • Figure 1: STL Partners’ AI Framework
  • Figure 2: Defining customer engagement categories
  • Figure 3: Customer experience is telcos’ main current priority with AI
  • Figure 4: The lifecycle of a chatbot/virtual assistant
  • Figure 5: Chatbots are a long way from meeting satisfaction levels of live agents
  • Figure 6: Telecoms operators say they lack the skills to deploy AI
  • Figure 7: Example of a rule-based versus AI-enabled chatbot in financial services
  • Figure 8: DT’s NLU and conversation flow selection and management process
  • Figure 9: Level of customer satisfaction with UK operators’ complaints resolution
  • Figure 10: Top reasons why customers complain to their service providers
  • Figure 11: A clear preference for predictive care
  • Figure 12: Estimate of telco opex breakdown

Big data analytics – Time to up the ante

Introduction

Recent years have seen an explosion in the amount of data being generated by people and devices, thanks to more advanced network infrastructure, widespread adoption of smartphones and related applications, and digital consumer services. With the expansion of the Internet of Things (IoT), the amount of data being captured, stored, searched and analysed will only continue to increase. Such is the volume and variety of the data that it is beyond traditional processing software and is therefore referred to as ‘big data’.

Big data is of a greater magnitude and variety than traditional data, it comes from multiple sources and can be comprised of various formats, generated, stored and utilised in batches and/or in real-time. There is much talk and discussion around big data and analytics and its potential in many sectors, including telecommunications. As Figure 1 shows, analysis of big data can give an improved basis upon which to base human-led and automated decisions by providing better insight and allowing greater understanding of the situation being addressed.

Figure 1: Using Big Data can result in richer data insights

Source: STL Partners

This report analyses how telcos are pursuing big data analytics, and how to be successful in this regard.  This report seeks to answer the following questions:

  • When does data become ‘big’ and why is it an important issue for telcos?
  • What is the current state of telco big data implementations?
  • Who is doing what in terms of intelligent use of data and analytics?
  • How can big data analytics improve internal operational efficiencies?
  • How can big data be used to improve the relationship between telcos and their customers?
  • Where are the greatest revenue opportunities for telcos to employ big data, e.g. B2B, B2C?
  • Which companies are leading the way in enabling telcos to successfully realise big data strategies?
  • What is required in terms of infrastructure, dedicated teams and partners for successful implementation?

This report discusses implementations of big data and examines how the market will develop as telco awareness, understanding and readiness to make use of big data improves.  It provides an overview of the opportunities and use cases that can be realised and recommends what telcos need to do to achieve these.

Contents:

  • Executive Summary
  • Big data analytics is important
  • …but it’s not a quick win
  • …it’s a strategic play that takes commitment
  • How is ‘big data analytics’ different from ‘analytics’?
  • Opportunities for telcos: typically internal then external
  • Market development and trends
  • Challenges and restrictions in practice
  • What makes a successful big data strategy?
  • Next steps
  • Introduction
  • Methodology
  • An overview of big data analytics
  • Volume, variety and velocity – plus veracity and value
  • The significance of big data for telcos and their future strategies
  • Market development and trends
  • Challenges and restrictions
  • Optimisation and efficiency versus data monetisation
  • Telcos’ big data ecosystem
  • Case studies and results 
  • Early results
  • Big data analytics use cases
  • Examples of internal use-cases
  • Examples of external use cases
  • Findings, conclusions and recommendations

Figures:

  • Figure 1: Using Big Data can result in richer data insights
  • Figure 2: The data-centric telco: infusing data to improve efficiency across functions
  • Figure 3: Options for telcos’ big data implementations
  • Figure 4: Telco’s big data partner ecosystem
  • Figure 5: The components of a telco-oriented big data

AI on the Smartphone: What telcos should do

Introduction

Following huge advances in machine learning and the falling cost of cloud storage over the last several years, artificial intelligence (AI) technologies are now affordable and accessible to almost any company. The next stage of the AI race is bringing neural networks to mobile devices. This will radically change the way people use smartphones, as voice assistants morph into proactive virtual assistants and augmented reality is integrated into everyday activities, in turn changing the way smartphones use telecoms networks.

Besides implications for data traffic, easy access to machine learning through APIs and software development kits gives telcos an opportunity to improve their smartphone apps, communications services, entertainment and financial services, by customising offers to individual customer preferences.

The leading consumer-facing AI developers – Google, Apple, Facebook and Amazon – are in an arms race to attract developers and partners to their platforms, in order to further refine their algorithms with more data on user behaviours. There may be opportunities for telcos to share their data with one of these players to develop better AI models, but any partnership must be carefully weighed, as all four AI players are eyeing up communications as a valuable addition to their arsenal.

In this report we explore how Google, Apple, Facebook and Amazon are adapting their AI models for smartphones, how this will change usage patterns and consumer expectations, and what this means for telcos. It is the first in a series of reports exploring what AI means for telcos and how they can leverage it to improve their services, network operations and customer experience.

Contents:

  • Executive Summary
  • Smartphones are the key to more personalised services
  • Implications for telcos
  • Introduction
  • Defining artificial intelligence
  • Moving AI from the cloud to smartphones
  • Why move AI to the smartphone?
  • How to move AI to the smartphone?
  • How much machine learning can smartphones really handle?
  • Our smartphones ‘know’ a lot about us
  • Smartphone sensors and the data they mine
  • What services will all this data power?
  • The privacy question – balancing on-device and the cloud
  • SWOT Analysis: Google, Apple, Facebook and Amazon
  • Implications for telcos

Figures:

  • Figure 1: How smartphones can use and improve AI models
  • Figure 2: Explaining artificial intelligence terminology
  • Figure 3: How machine learning algorithms see images
  • Figure 4: How smartphones can use and improve AI models
  • Figure 5: Google Translate works in real-time through smartphone cameras
  • Figure 6: Google Lens in action
  • Figure 7: AR applications of Facebook’s image segmentation technology
  • Figure 8: Comparison of the leading voice assistants
  • Figure 9: Explanation of Federated Learning