The need for transformation has been bought into sharper focus by the pandemic – with employees working from home and customer purchasing online pushing forward, personalization, cross-channel customer understanding and new intelligent tools to optimize their business and operations.
The Prioritization of Intelligence In Telcos
Technologies such as AI and ML offers pattern recognition and anomaly detection in many areas across the telco – able to deal with large volumes of data and make complex, accurate decisions at a speed. They can also provide human-like actions such as the ability to recognize images or understand speech. Telcos usually prioritize intelligence projects by linking them with major corporate-level goals. Including, providing good connectivity, quality services and improving customer satisfaction with capabilities such personalization or more proactive decision making to avoid or rectify customer-affecting problems. Into the longer term, the automation of the network and major processes will enable telcos to lower their cost base and support new revenues from more complex services like IoT. Calculating the total potential financial upside from implementing intelligence, a “typical” telco generating $15bn in revenues a year can expect around $1.2bn of upside from intelligence projects in the next 5 years.
Drivers of Intelligence Deployment
Deployment of successful AI/ML depends on a number of activities within the telco: they need to make the technology work and projects can underperform if the training data is not of good quality or the algorithm doesn’t produce clear and usable results. Also, in order to decide on the action to be taken, more complex intelligence is needed which will takes time to develop and optimize. There also needs to be orchestration capabilities in place to carry out the action, requiring the roll out of technologies such as SDN/NFV. Lastly, there remains the underlying issue of being ready organizationally for the addition of intelligence: Do staff have faith that the intelligence will work? Does the telco have the right skills to create, manage and work with it? Is there senior management buy-in and a vision for what could be created in future?
Key Areas for Deployment of Intelligence
Although, telcos have their own data science teams, a large percentage of the $1.2bn upside will come from either packaged intelligence products where the vendor brings specific expertise, or from products with intelligence embedded. The telco will purchase this intelligence in a number of different ways:
- Some intelligence for the network will be purchased as part of new management systems but SDN/NFV provide new opportunities for specialist vendors to package solutions to address particular needs.
- The Chief Data Officer (CDO) has responsibility for the use of data across the telco, dealing with data management, cultural transformation and governance issues. They will have data science teams under them, creating individual algorithms, typically, for marketing, sales and some specific network issues.
- There may also be a CDO (or similar job title) within the enterprise team of larger telcos. Focused on B2B data problems such as sales and pricing; and responsible for the monetization of data externally. They will have data teams under them creating their own algorithms, however, they often require specialist data products to enable monetization across a range of verticals.
- Lastly, there are multiple types of intelligence which will be embedded into department-specific technology – for example, the contact centre, supply chain, finance or HR.
Into the future, telcos will have three main options: building their own intelligence, purchase intelligence as part of a complete solution or integrating an intelligence product into existing or new technology. As vendor intelligence solutions mature, the decision to “build” is likely to be dependent on existing capabilities in the telco’s data science team, seeing the team increasingly focussing on a small range of areas. Purchasing an integrated solution will be preferable where the vendor can use its intimate knowledge of the process/the underlying data and expertise from working with a range of telcos – and is likely to the preferred option in many situations where the telco needs to purchase additional intelligence.