How Liberty Global approaches GenAI adoption

During a fireside chat at Network X in Paris, Liberty Global's vice president for data strategy and AI, Nirali Patel, shed light on the company's approach to implementing AI across its organization.

Tereza Krásová, Associate Editor

October 11, 2024

3 Min Read
Liberty Global's Nirali Patel (left) and Omdia's Ronan de Renesse on stage at Network X
Liberty Global's Nirali Patel (left) and Omdia's Ronan de Renesse on stage at Network X(Source: Tereza Krásová/Light Reading)

Generative AI (GenAI) remains a hot topic in the telecom industry, with companies looking at ways it can be implemented to streamline operations and offer a better service. Liberty Global's vice president for data strategy and AI, Nirali Patel, discussed the company's approach to implementing AI during a fireside chat at this week's Network X show in Paris, which was moderated by Omdia's vice president for telecom research, Ronan de Renesse.

The company has already implemented AI in a number of ways, Patel noted, including to predict faults, churn and to create recommendations on media boxes.

For example, Liberty Global can use AI in fiber deployments to determine what each engineer is good at, and even what they like doing, and then assign certain tasks to the appropriate person. This, said Patel, led to more jobs being completed per day.

Meanwhile, GenAI has brought significant benefits for the company, she said. One is to make data more accessible by processing unstructured data from sources like voice calls, field engineer notes or social media to get a better picture of what customers want.

Liberty Global, Patel said, opted to focus on a limited number of areas to implement GenAI, choosing three streams for its adoption: call center agents, engineering workforce and internal employees.

Consumers of connectivity services, said Patel, care little about the underlying technology, making it important for Liberty Global to bring different silos together. She pointed to a solution released in several of Liberty Global's markets that listens in on conversations and can suggest more proactive approaches to solve a problem. This is done by bringing together different datasets like network systems, CRM and billing systems.

Asked how Liberty Global evaluates potential new AI initiatives, Patel said the first step is to look at feasibility. This includes assessing the availability of data, regulatory issues and privacy concerns, as well as cost. She also stressed that feasibility also includes change management feasibility: "Have we got the right mindset, the buy-in and the capacity as an organization in this industry to actually embrace this solution and make it a long-lasting change?"

Make it or buy it?

De Renesse also asked Patel how Liberty decides whether to partner with others or do things in-house. Examining the risks involved is a key factor, she said, adding that buying a solution or bringing in a partner to help build it can be the better option.

She also stressed the importance of having an agile mindset and being able to step away from developing something in-house if a better alternative emerges elsewhere.

She cited the example of HR assistance agents, saying that many companies have decided to build one themselves using GenAI. Liberty instead opted to watch and learn, subsequently finding that suitable agents were already available on the market.

To help it with AI implementation, Liberty Global has assembled a common AI team while also partnering with companies including consultancies and solution providers. The approach has helped with the deployment of products, said Patel.

Even so, while individuals have become used to mainstream AI products like Siri or Google Translate, introducing AI in organizations is much more complicated, said Patel. She stressed data quality is a key factor, pointing to the "garbage in, garbage out" concept, although she did note the industry has improved in this area.

Read more about:

AIEurope

About the Author

Tereza Krásová

Associate Editor, Light Reading

Associate Editor, Light Reading

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like