Nirali Patel, Group Vice President of Data and AI at Liberty Global, talked to TM Forum Inform about the group’s new strategy and why now is the “telco era” for AI.
Liberty Global “dream team” drives group-wide AI strategy
Generative AI (GenAI) has broadened the opportunities for telcos to reduce costs by becoming more efficient and to grow revenue by leveraging their data. As the sector grapples with the pace of change, Nirali Patel, Group Vice President of Data and AI at Liberty Global, has created a multi-disciplinary “dream team” to deliver a common approach to AI across the group.
Speaking to TM Forum Inform from Liberty Global’s London headquarters, Patel’s enthusiasm was palpable, even over a video call. She said GenAI is a “step change” for Liberty Global and the telecom sector in general.
“We had been using AI, but now suddenly everybody could use it. There’s been an explosion of AI ideas … And there’s a whole universe of data that’s accessible to us now through Gen AI, which means there are more opportunities for us to either become more efficient or generate revenue,” she said.
“It’s a telco era now” because they have a “massive opportunity” to improve operational efficiency and monetize their data, she explained.
Before GenAI tools hit the scene in late 2022, telco data and AI strategies were less mature and smaller scale than in other sectors, in her view. Prior to joining Liberty Global in January 2023, Patel worked in healthcare and financial services and served as the Chief Data and Analytics Officer at BT’s Openreach till the end of 2022.
Liberty’s strategy aims to ensure a collaborative approach across the group “so we can have scaled AI products” and “to be selective about which areas we focus our energies on.” Rather than having hundreds of AI applications, it is better to have a few done well and used by all operating businesses, she explained.
Liberty has prioritized AI use cases in three areas: “customer, network and employee”. Broadly, the operator is developing AI solutions that can improve customer experience and network quality and make employees’ jobs easier.
Consistent and separate
It is a unique challenge to establish a common approach at Liberty Global because it is a federated organisation with independent cable, fiber and mobile operations in six European operators. The group’s operating companies are Sunrise (based in Switzerland), Telenet (Belgium), UPC Slovakia, Virgin Media (Ireland), Virgin Media O2 (UK joint venture with Telefónica) and VodafoneZiggo (joint venture with Vodafone in the Netherlands).
To achieve as much consistency as possible, Patel said 60% to 80% of AI products have common technology and the rest is added by the local operators to tailor offerings to their specific markets.
Another complexity for the development process is the need for separability, she said. That is, products are designed so that they can be decoupled from the group if Liberty acquires or sells holdings in operating businesses. For example, Liberty Global plans to spin off 100% of Sunrise in the fourth quarter of this year.
“When we develop our AI products, I have to create them in a way that I can still provide service to a market that may no longer be part of our family … through a service level agreement process. I also create products that I can deploy in an easy way into a new business when it joins us,” she said.
Dream team goes for goal
Patel has created what she describes as a dream team to deliver AI projects and meet the group’s objectives. The team comprises people from Liberty Global, the operating businesses, and partners who come together in various combinations on projects to develop, scale and integrate solutions. They tackle aspects, including the AI engine, data management, IT development and integration and change management.
“When we’re delivering these products, it’s an end-to-end solution. I’m trying to minimize what the market has to do post that point,” she said.
Assembling a multi-disciplinary team in this way shows Liberty’s efforts to bring consistency to the group of operating businesses as well as raise awareness among employees.
Patel said it was also important to set clear goals for people and culture. “It’s not just about technology,” she said.
“We want to make sure that we equip all our leaders to understand AI…ensure that our employees have the right skills and attitudes towards AI … and prepare them for the changes to the business processes that they’ve been used to,” she said.
Given that the AI strategy touches every part of the organization, implementing it has been a “game changer” for increasing collaboration.
“We are working together way more across our business units to get the successful outcomes that we want out of this,” said Patel.
An AI assist for customer service agents
One of Liberty’s recent initiatives is an AI platform that supports call center agents, called Agent Assist. The platform can predict why a customer might be calling based on their call history, and it automatically generates a call transcript and summary, which has been a manual task for agents. It also provides customer insights for the operator.
Agent Assist has been deployed at VodafoneZiggo in the Netherlands, where 200 agents are using it. Initial results show that the technology has reduced overall call handling time by 5%, while also receiving positive feedback from customers about their engagement with the agents.
“It has saved our agents time and effort, but also they feel that they can focus on the customer more,” said Patel.
Virgin Media O2 is also testing similar AI technology for its broadband customers in the UK.
Another high-profile example is Sunrise’s use of AI in mobile networks to save energy, which was implemented in 2021. The technology reduced Sunrise’s mobile network energy use by 10%.
Is AI for cost savings or revenue?
Telcos are often asked if AI strategies are aimed at cost savings or revenue generation. For Patel, it’s both and they are related.
With Agent Assist, the AI platform saves time for agents but it also delivers information that can be used to develop new products.
“It gives us a whole new data set that we can get insight from … that could lead us to create a new product or change one. Until now, we haven’t been able to do that at scale,” she said.
The platform provides data from tens of thousands of calls that can tell Liberty’s operators what products need to be changed, improved, or even stopped.
“There's only so much value our business is going to get out of efficiency. We're going to get more value and success out of new revenue generation and differentiation of our products,” she said.
That’s why she believes AI is “game-changing” for telcos.
“It's really important for telcos to figure out where they can differentiate. When you move towards the goal of having autonomous networks, where you're using AI, autonomy and robotics to make your network zero touch, your costs reduce…but your profit just plateaus. To survive in the industry, you need to have some other things that you can sell … to keep growing otherwise you’re not going to grow as a business,” she said.