Telia supercharges its data with gen AI

Data is the backbone of telecom operations, driving superior customer service and informed decision making. But according to Accenture's "From Survive to Thrive" report, only 28% of senior executives of the companies surveyed feel confident in their ability to provide easy, decentralized data access.

Telia, a major telecom provider in the Nordics and the region’s internet bedrock, knew that becoming a data-led organization would be key to unlocking productivity and growth. Yet, in one of their major data warehouses, only 1% of its data had corresponding metadata, making it challenging to find, understand and act upon crucial information like call records and customer interactions, including sales and service inquiries. This presented an opportunity for Telia to enhance its data utilization and improve its ability to analyze and make informed decisions.

Telia partnered with Accenture to create a generative AI solution that automatically updates the telecom’s vast data repository with metadata descriptions. We used our AI and Amazon Web Services (AWS) expertise to leverage Amazon Bedrock, building an intelligent solution that updated tags and descriptions for 87,000 data fields, saving Telia 10,000 data engineer hours and €700,000 in data entry costs. To illustrate the scale of the work, think of the company’s data as a vast library with thousands of unlabeled books. Previously, employees had to sift through these books without clear labels or summaries. Now, with the AI-powered solution, it's like having dedicated librarians who meticulously update the labels and summaries for each book, making it easy and quick to find the right information.

This solution is a crucial step in enabling Telia to become a data-driven organization.

Amplifying a data tune-up

Telia’s journey to democratizing data and becoming truly data-driven went beyond saving time and money — it set out to spread data literacy across the entire company and enable the organization to leverage data for insights and smarter decisions. For that to happen, its data storage model needed to be simple enough to grant easy access, and user-friendly enough so that anyone sourcing the data could understand it.

Accenture used Amazon Bedrock to analyze Telia’s data and automatically populate corresponding metadata descriptions. By providing the AI model with prompts and examples of data descriptions, it was able to generate precise and detailed metadata for columns that were missing data. The descriptions were found to be more accurate and efficiently created by AI than those created manually. With a centralized and updated metadata repository, Telia could also improve its data governance and ensure all assets were properly managed and compliant with regulations. This has further enhanced the data security and privacy of confidential customer data.

Telia’s data experts are essential to maintaining the solution. It relies on a ‘human-in-the-loop’ review system that flags AI-generated column descriptions, so that Telia’s people know which data to double-check. We helped provide change management and training services for Telia’s team around how the solution works, how to maintain it effectively, and how to measure its accuracy. This comprehensive approach has helped Telia become far more self-sufficient and data-driven.