Generative AI (GenAI) has been in headlines non-stop since the beginning of 2023 and the rapid adoption of ChatGPT by the consumer market.
Since then there has been a massive rush by the technology industry to innovate, iterate and develop commercial use cases for GenAI across all industries. One industry that is a clear target for the evolving Generative AI ecosystem is telco.
A range of different companies from the telecoms ecosystem, with different competencies and product and service profiles, have recently developed Gen AI service offerings dedicated to supporting telecoms operators. This comes as the market is moving towards the development of industry specific models to support more specific use cases, often using what's termed a Small Language Model (SML) as opposed to the large language models (LLMs) that drove the original hype around Gen AI technologies.
There is also wider development of model distillation to create smaller and more efficient models based on LLMs.
OpenAI introduced Model Distillation in October 2024 while AWS announced Amazon Bedrock Model Distillation recently. These smaller models run less parameters than LLMs reducing the cost to run inferencing, and are often supported with additional enterprise specific datasets to provide deeper insights into domain specific telecoms data.
Emerging players looking to support Telco specific GenAI services range from the telecom equipment vendors, global systems integrators, OSS/BSS, hyperscalers, and other telco software vendors as well as some of the telcos themselves. Each emerging player is approaching the market from a slightly different value proposition, although the early uses cases targeted by these players have quite a lot of overlap.
For example, network vendors such as Huawei or Nokia have developed telco specific models which are fed data from the underlying infrastructure to create network digital twins as well as AI-agents that can be interfaced with to help support network operations and management teams with natural language ways to query about issues within the network.
Meanwhile, OSS/BSS providers like Amdocs or NetCracker are leveraging their position in providing critical software stacks to telcos, and have built out GenAI platforms that include industry standard LLMs, as well as their own telco specific LLMs to support operators adoption of GenAI tools in areas like billing support, customer on-boarding, service assurance, network optimisation and more.
Global Systems Integrators, which have been among the most active company types in developing commercial grade generative AI services for industry, have also set their sites on the telco vertical. Companies like Infosys and TechMahindra have built their own AI platforms incorporating multiple LLMs, including third party and proprietary. They are now targeting telco with domain specific models that touch on network troubleshooting, customer service and network design.