AI: Last chance saloon for non-connectivity aspirations
Daniele Pe - Senior Principal, San Francisco office
Andrew Snead - Managing Partner, New York office
The rise of AI and its adoption across corporates
Although the potential of artificial intelligence (AI) has been discussed for the past 50 years or so, we’re only now starting to witness meaningful progress. Such progress is being driven by advances in machine learning (ML), a subfield of AI, and more specifically deep learning, a branch of ML that roughly imitates the structure of the human brain (i.e. virtual neurons) to process data and recognise more complex patterns. These advancements, coupled with several ecosystem factors – dramatic growth in data availability, improvements in computing power at lower cost and increased availability of VC funding – have created the foundations and ingredients for AI to take off.
An inordinate number of articles are written daily on how AI will transform our world, from finding cures for non-communicable diseases to displacing significant parts of the labour force with robots and virtual workforces. The truth is, we simply do not know yet. The applications and use cases will extend far beyond what we envisage today and the implications, both positive and negative, will likely change the fabric of society. The manner in how we interact, communicate, consume and buy will change significantly which is why global Tech giants (Amazon, Google, Apple, Microsoft and Facebook) and VCs alike are investing significant capital and resources into its development.
These companies have all made several AI investments in the last 3 years and dominated AI related transactions. Given their financial resources, software heritage and relentless focus on innovation, it is not surprising these Tech giants are leading the AI game, but it is surprising how corporates in other sectors are failing to respond, perhaps waiting for the hype to settle (or fearing another transformation fail, scarred from the past), but in doing so, putting their future at risk.
For reasons we’ll explain later, the Telco sector is one that stands to be particularly impacted by AI, yet few operators are investing and exploring its relevance in a meaningful manner. Examples include the investment arms of NTT docomo and Singtel placing bets in emerging AI players; SK Telecom launching a sort of Amazon Echo-like home assistant, and similarly Telefonica announcing its own AI driven virtual assistant to manage customer service interactions; as well as AT&T releasing its AI open source platform to simplify AI deployments and lower entry barriers for companies wanting to explore it. However, most operators have yet to embrace the topic, instead focused on eking out an extra percentage point of revenue from their core connectivity businesses.
Naysayers will point out that AI is far from a mature technology. They are right. AI agents need to be explainable and trustable whilst interaction needs to shift to something far more conversational. However, it’s hard to believe such issues won’t be resolved in the near-term, given the capital and resources being deployed by the large tech firms.
So, what’s the problem?
The problem is AI is set to redefine the entire Telco model, which means those Telcos that are not embracing its relevance (and aligning investment and management focus accordingly) are set for an uncertain future. Industry observers have been writing about Net-Co / Serve-Co for years. The reality is that ‘Serve-Co’ will likely mean ‘Tech-Co’ as AI-fuelled virtual assistants ultimately replace the need for mass market propositions, pricing plans, high-street shops and call centres. Ultimately, AI will enable the dynamic provision of tailored, model-of-one propositions to individuals in line with their specific needs, location and context (which will be predicted and planned for in advance) whilst the underlying resources will be adapted accordingly. The entire model will change. Who is best place to do this? A Tech-firm with bootstrapped access or a Telco with bootstrapped Tech?
Let’s unpack this further. AI can fundamentally transform the Telecom space in three main ways:
1. Make Telecom companies’ operating model and infrastructure layer significantly more efficient
AI can be used to revolutionise the way customer service is delivered and internal processes are managed. As mentioned above, Telcos such as Telefonica and Orange are starting to use basic forms of AI to replace customer interaction and detect fraud, but this is only the tip of the iceberg. AI can be used across any department with functionalities such as personalised product suggestions, pricing and promotions optimisation, customer satisfaction monitoring and prediction, virtual sales and customer service support, automated inventory management, CV screening and hiring recommendations, working capital optimisation, to name a few.
Exhibit 1. AI applications across Telecom operators’ functions
Beyond operating model and key processes, AI will have several applications at the network layer. In terms of network design, AI will be leveraged to further optimise roll-out through better understanding of network usage and capacity forecasts, and will limit additional deployments by enabling self-adjusting features that optimise coverage and utilisation. In terms of network maintenance, AI software agents in the network will continuously compute and predict the future reliability of the network, the failures and the network degradation whilst scheduling preventive maintenance in an optimized manner with zero human touch. Furthermore, AI will help detect patterns for cybersecurity purposes and address them around the clock.
Those activities will drive significant P&L benefits: higher revenues thanks to better cross-selling and optimised pricing, as well as OPEX and CAPEX reduction through automation, improved efficiencies and elimination of redundancies. Such efforts will also enhance overall operator customer experience.
2. Reduce the entry barriers for the Tech-centric MVNO
As AI matures, and other technologies such as blockchain, robotics and IoT evolve, we expect to see the emergence of a new breed of software-centric Mobile Virtual Network Operators (MVNOs). Their opportunity will be to exploit the lethargy and sluggishness of traditional incumbents, deploying AI enabled business models sooner and with greater effectiveness (as well as greater disruption levels), taking full advantage of the benefits. The advent of soft-SIM coupled with the emergence of urban-centric Net-cos (utilizing Wi-Fi and unlicensed spectrum) will act as additional enablers. The threat to traditional players could be considerable.
3. Threaten the efforts that operators have been making in creating digital services hubs
In the last few years, operators have – with differing levels of success - collaborated with digital players as an attempt to create a ‘digital life’ experience, aggregating several digital services into a cohesive offer with varying degrees of sophistication, ranging from the more basic T-Mobile ‘Binge On’ offer to the more sophisticated communication and entertainment app recently launched by VEON.
Yet, it is by looking into at Digi’s (Telenor Malaysia) “Music freedom” offer that one can really understand how AI can disrupt these operators’ efforts. “Music freedom” aggregates over 30 music streaming services in a new mobile application, providing additional features such as exclusive content, contests, tickets, as well as an overarching layer with an overview of top songs across these music services and easy navigation between them through deep linking. The objective of this product, beyond helping Digi position itself more as a content player, is to provide customer with convenience – i.e. help customers overcome the fragmentation of mobile music streaming. This aggregation layer role becomes superfluous with AI. In fact, AI is able to help customers navigate through the plethora of digital services and provide the best service for their needs without the need for a specialised app/hub to do that, as evidenced by the increasing array of services being offered by Google Home and Amazon Alexa.
Exhibit 2. Existing virtual assistant examples
As the plethora of integrated services expands, and conversational AI becomes a reality, customers will shift their behavioural patterns in terms of the way they request, select and consume. The end of the smartphone? Maybe, at least as we know it. Irrespective, it’s clear much of the customer engagement and interaction will happen independently of the smartphone and independently of operator digital hubs, at least in their current incarnation.
The likelihood of Telcos participating successfully in this space is low. Even if they would be able to acquire or develop the AI capabilities required to deliver virtual assistant services, there is the question of time. ML is an iterative technology that continues to learn and improve as it consumes, digests and makes use of unstructured data. Amazon and Google are already harnessing their vast quantities of customer data to train and evolve their AI capabilities which in turn, is improving the tailoring and attractiveness of their products and experiences, creating stickiness and subsequently more user data. A virtuous circle that makes it ever more challenging for traditional players (and many other corporates) to catch up and compete. The answer perhaps lies in an integrated solution – an assistant within an assistant.
What does this mean for Telecom operators?
The combination of these factors poses some fundamental questions for traditional players. The role of the Telco will further diminish in terms of customer engagement and end-user relevance. Connectivity will remain an attractive business so long as operators can improve their return on invested capital (ROIC), given its importance to shareholder value. This will increasingly require scale economics, smarter CAPEX deployment and alternative network models. The challenge for many Telcos is they aspire (realistically or otherwise) to participate beyond the connectivity layer.
So, is the battle completely lost or can Telcos change the course of history?
Whilst difficult to accept, the non-connectivity part of the telecom business as we know it is dead. Operators have consistently failed to innovate and have time and again missed the opportunity to leverage their unique assets, particularly customer data, to position themselves strongly in the digital services scene. How the industry lost messaging will be debated in business school classrooms for years to come (along with banking and payments). The ‘Kodak moment’ however, perhaps lies in wait.
AI is potentially both the final straw and opportunity, at least for those aspiring to participate beyond connectivity. The question is what can Telcos do to give themselves a realistic chance of harnessing not just AI but other technologies to truly transform their business in the form of a ‘Tech-Co/Net-Co’.
Some (non-exhaustive) potential options:
OPTION 1: Stick to the knitting, avoid wasting shareholder money and rather enable those who can
This course of action seeks to enable software-centric MVNOs / Tech players and shifts the focus away from selling digital services to end-users. It broadly accepts the battle for end-user relevance (beyond connectivity) is lost and seeks to generate value between the pure connectivity and digital services layer. It does not try to build an AI informed / software centric Tech-Co. The primary focus would centre on selling B2B2X platform services to a range of players, likely to include the longer tail of Tech players and MVNOs but also other ecosystem players (see Exhibit 3). This requires productising the key assets via horizontal enabling capabilities (cloud, security, authentication, analytics etc.). Investment profile would be moderate yet capabilities (either via partnerships or targeted acquisitions) will be necessary in specific areas. Finally, this model necessitates a very lean operating model and cost structure.
Exhibit 3. Conceptualisation of the Telco as a platform player
OPTION 2: Accept the elephant will not dance and build a Tech-Co by the side
This option accepts the likelihood of successfully transforming a traditional operator into an agile software-centric Tech-co is low and rather seeks to solve the problem by acquiring / building a Tech-Co by the side, via a conglomerate holding structure. This option assumes synergies can be generated from sufficient collaboration and capability sharing between Tech-Co and Net-Co whilst also allowing Hold-Co to monetize Tech-Co assets and investment upsides, beyond the link to Net-Co.
OPTION 3: Transformation into Tech-Co – the bold and the beautiful (and likely fail?)
This option assumes the operator fully commits to transforming itself into a Tech-Co and placing AI and other Technologies at the centre of its business. It’s radical and would require full shareholder support which, itself is challenging, given the institutional nature of most Telco shareholders and their quest for dividend certainty. It would require some bold and sizable investments and a restructuring of the company (to convince new talent of the seriousness of intent) with a clear Tech-Co anatomical set-up, DNA and culture.
While still providing connectivity services, the end-user proposition would be radically different, blurring the lines between telecom and Tech: a fully app-centric environment interacting with customers through a multitude of customer interfaces – from voice and gesture recognition to screen visualisation via AR/ VR providing a variety of services: from hyper-curated content and other digital services experiences, to data brokering and privacy management. All services would be fully personalized and dynamic, managed by an omnipresent digital assistant who ultimately operates on behalf of the user (in a quasi-digital twin capacity). No need for contracts, tariff plans, rigid propositions, shops, customer care centres.
Exhibit 4. Conceptualisation of the operator of the future
The obvious challenges are twofold: likelihood of full shareholder support, given the point outlined above, and likelihood of executing successfully. Based on the industry’s track record, many attempts are likely to fail.
While aspects of these models might seem far-fetched, there are examples of operators that have started moving in such directions.
Softbank is pursuing the diversified TMT holding route, transforming itself into a diversified investment company, with a Group holding that owns telecom assets (Softbank, Sprint), major Tech assets (Alibaba, Yahoo Japan, Brightstar, ARM) as well as several Tech funds, such as the $100BN vision fund.
VEON has repeatedly announced their intention to become a Tech-co, even touting a potential exit from the network infrastructure business. They recently launched a digital services app, taking on WhatsApp (a gutsy move) and using customer data (guided by user preferences and permission) to curate 3rd party services.
SK Telecom recently announced its ambition to transform itself into a holding company, closer to an Alphabet than a telecom operator, to increase the relevance on their already significant digital services platform (SK Platform).
However, it is also likely such an outcome could be the product of a reverse move by an existing Tech company. i.e. global Tech-co buys an operator (or acquires spectrum) and builds such a business, bootstrapping access to its Tech capabilities. The “will they, won’t they” question surrounding Google (following the Google Fi launch) and Line (prior to and following the launch of their MVNO in Japan and more recently in Thailand) has been discussed at length. Many argue it makes little sense for such players to enter the Telco world (why would they want to?) yet their need to further innovate and push boundaries may require them to do so, more out of necessity than desire.
Time will tell. What is most critical for Telcos is to make a deliberate choice and fully commit to whatever path they decide to take. Unfortunately, strategies and ambitions change with CEOs and management teams get too easily convinced they are on the right path, blinded by isolated initiatives that are grossly insufficient to make a meaningful difference. For those seeking to participate beyond the connectivity layer, this may well be their last chance: last chance saloon.
© 2020 Delta Partners.