Inês Matos, Head of Network Data Science & AI at Vodafone and Celfocus’ André Vieira discuss the business case, real-life implementation and the operator’s ongoing AI strategy
Inês Matos, Head of Network Data Science & AI at Vodafone Vodafone and André Vieira, Pre-sales Senior Manager at Celfocus held a joint session to discuss – and provide real-life experience – of moving agentic AI from theory to business practice. They were speaking at Mobile Europe‘s recent Telco to techco virtual conference.
WATCH THE VIDEO ON DEMAND HERE
Veira set the scene, describing agentic AI as an evolution of GenAI. He explained, “First, we had the the LLMs andthey were great. Everyone was using them when ChatGPT launched, but they were not suited for the enterprise. They didn’t have the full functionalities that we needed, that then we had the evolution of the RAGs [Retrieval Augmented Generation] talking with private data, but still missing the part was automation…we see the AI agent as the automation, the chatbot for automation and the orchestration of these automations. So we [agentic AI] as little helpers for the engineers.”
Leveraging cloud, avoiding lock-in
Matos then explained that Vodafone’s AI strategy is hybrid (see below); it will continue to leverage existing automations such as use traditional machine learning, and to complement them. To this end, Vodafone created an AI framework.
Although Google Cloud is Vodafone’s strategic partner, the operator is wary of vendor lock-in so has also created a gateway to Microsoft and Matos said, “In future, we will also be open to AWS. And this is how we are planning on growing. So we already have systems that are working. We want to put GenAI on top, we want these agents to come on top to help with the network performance…and through the network performance and helping our developers, we want to drive sustainable growth. And we want to optimise the networks.”
Three-strands of automation
She continued, “We have defined three big bets. So one of them, is our zero-touch ambition on the networks which will eventually get there. The second one is the network investments – we want to be smarter [about] where we invest our money, and where it has the most impact in the network. And the third one is what we call Vina [Vodafone Intelligence Network Assistant]. So what happens with this assistant is that this is
Bringing them together
Matos says VIna is the “GenAI part where we want to create an umbrella…connecting to different agents. Each agent is…coming up as different proof of concept from different teams.” She provides some fascinating detail about the work the teams are doing, and why, in creating independent agents.
Then, as Matos explains, “Vina basically integrates all these assistants so that they are not confined to one team that knows their own problem very well – you can now get a bird’s eye view of the different type of data [generating] different documentation and basically crossing this data to to help a better view of the network.”
Find out more about what Vodafone is doing, in detail, and future plans plus Celfocus’ input from the on-demand video here.
