.gif?lu=913)
.gif?lu=925)
27 February 2025
Global telcos are eager for generative AI solutions but are struggling to make them work, with only half of use cases meeting business goals, according to a new report from Altman Solon and Kyndryl.
The report, which included a survey of more than 100 global telco leaders, shows that 85% of telcos have at least one generative AI use case in production but only 30% of all use cases are making it to production. Generative AI has also proven to be more expensive than expected, with 80% of respondents reporting higher than budgeted implementations.
“The potential for gen AI has so far exceeded the reality for global telecommunications providers, but it’s very early in the development cycle for this promising technology,” said Altman Solon Associate Partner Priya Mehra. “We are already seeing some telcos adapt with cost management solutions and stronger AI governance to deliver better results.”
“One of the key lessons of the report is that a ‘one-size fits all’ approach to generative AI does not work in the telco space,” said Altman Solon Partner Diane Leung. “Recognizing – and addressing – pressure points and leveraging third-party experts can help telcos sustainably scale gen AI implementations.”
Other findings include that 50% of respondents set a minimum threshold of 10% ROI in their business case to move their Gen AI use cases from pilot to production. Proficiency in data management and governance is directly tied to realizing success and stated business objective from Gen AI implementations. Four archetypes of telco organizations have emerged, with two, ‘Broad Experimenters’ and ‘Cautious & Focused,’ achieving higher success rates. Telcos are evaluating model deployment choices and using multi-LLM strategies within workflows to manage costs.
