By Kamesh Chelluri and Narendran Sivakumar

Fast broadband and mobile data networks have propelled the ‘app revolution’. But the revenue streams have largely bypassed the operators who built them in the first place. 

To avoid their services becoming a commodity and networks ‘dumb pipes’, telcos must reinvent themselves. Whether they decide to offer smart connectivity, create a portfolio of smart services, become content aggregators or even content creators, one thing will be essential to their success: they need to become ‘cognitive telcos’. 

This means imbuing their networks and services with automated, data-driven intelligence as competitive ammunition, to reassert themselves within the value chain and capture a larger piece of the revenue pie. 

The ‘app revolution’ revenue streams have largely bypassed telcos. (Source: Shutterstock)

Becoming a cognitive telco

The telecommunications sector has been an enthusiastic adopter of automation, powered by Artificial Intelligence (AI)-based technologies from early on. In 2017, McKinsey found telecoms had the highest proportion of companies which had adopted three or more AI-related technologies at scale.

The reason is obvious. Telecom operators have always had huge amounts of data on which to unleash AI algorithms and machine learning (ML): customer data, network data, usage data, financial data, and so on.

What’s often still holding them back is the siloed nature of these vast data repositories, which are typically spread across many internal departments with little to no overlap. Operators need to cut across such division and make data available comprehensively to the entire organization to develop effective future strategies and products. 

Yet it’s is only a preliminary step toward becoming a ‘cognitive enterprise’ – a firm which makes use of joined-up intelligent systems that can sense, infer and reason, and respond to data. 

One step at a time

The first stage – process automation – is now prevalent in many industries. Algorithms replicate human tasks such as order entry, basic event monitoring or data capture.

The second stage – reactive autonomy – involves software that can assimilate data and act on its basis. Examples include processes such as compliance checks or ‘onboarding’ new subscribers, which can be programmed and executed automatically. 

The next step – proactive autonomy – looks at the business in real-time, for instance to predict ‘churn’ and take pre-emptive action to retain customers at risk of leaving. Other examples include identifying and fixing order failures, and predictive maintenance, which calculates when faults may occur and nips them in the bud. This level of automation is already quite widely – and successfully – used among network operators. 

Intelligent AI systems could make error messages like these a thing of the past. (Source: Shutterstock)

Prescriptive autonomy is one step up from this, and involves algorithms making deductions from data and responding to them. Many of today’s media enterprises use this level of AI to understand customer preferences and to personalize services. For instance, your streaming service can analyse which TV shows you like, how and when you consume them, and provide targeted suggestions without any human intervention. 

Beyond that lies the final stage – cognitive autonomy. This occurs when data and intelligence is shared across the entire company, giving a 360-degree view of every operation and interaction. At this level, AI-enabled systems can sense a network event, identify its root cause, deduce the implications and come up with a response. 

Take the following scenario: a customer decides they want to upgrade to a new phone. This request could trigger a personalized sales offer, with a tailored package based on previous behaviour and consumption patterns. It might also include new services based on the technical capabilities of the new phone. Once the subscriber has signed up, all adjustments to the network take place automatically. Their decision will then also feed into refining future sales offers, new product development and even the partners a telco may want to work with. 

All this takes place automatically, tying in all relevant systems, teams and departments. This is also referred to as ‘zero touch’. The result is a sustainable, self-perpetuating model that continually draws on information to optimize business and network operations. 

Where next?

The opportunities for cognitive telcos are substantial. But maximizing this potential means thinking big. In creating their roadmap, CEOs and CXOs should be asking what they need to do to institutionalize an operating model for data, considering the implications for organizational structure, people strategy, governance and business values. Such a holistic approach will ensure that raising up the business in the communications value chain is at the heart of the transformation, not the technology.