Telecoms operators’ approaches to GenAI solutions should be shaped by priorities, experience and budget

Operators are progressing with their adoption of GenAI. In the first few months of 2024, some operators rolled out new GenAI-based services; some of which were created internally and others via partnerships with vendors. Other operators are investing to enhance existing data centres with AI cloud infrastructure so they can create enterprise services that are based on GenAI and generate new revenue. Despite this progress, operators are facing challenges with GenAI deployments, including the complexity associated with using GenAI capabilities from multiple environments. Operators need to adopt a GenAI platform that can accommodate a common set of tools to support the lifecycle of foundation models/large language models (FMs/LLMs) and planned GenAI use cases. Operators will also need to develop a clear view on how to implement this platform.

This article summarises the three GenAI platform implementation strategies covered in our recent report on GenAI and the telecoms sector and outlines how operators can decide which approach to adopt.

As operators advance with GenAI, they should consider adopting a GenAI platform to facilitate and fully deploy GenAI solutions

Operators are continuing to deploy GenAI solutions and are engaging with vendors to extend existing AI platforms to support GenAI-related development. At MWC 2024, for example, operators such as Deutsche Telekom and KT launched ‘ready-to-consume’ GenAI solutions for their customers to improve employee productivity and customer experience. Maxis Telecom established partnerships with public cloud providers (PCPs) including Amazon Web Services (AWS) and Google Cloud to leverage their AI cloud infrastructure and foundation models/large language models (FMs/LLMs), in order to develop applications for internal operations and revenue-generating services for enterprise customers. Meanwhile, Rakuten Mobile and Telefónica announced collaborations with OpenAI and Microsoft Azure respectively, to integrate these vendors’ GenAI capabilities with the operators’ pre-existing AI development environments.

These activities indicate operators’ growing interest in GenAI and the increasing maturity of their journeys. However, they also highlight the urgent need for operators to assess how they deploy and manage their GenAI use cases to ensure consistency and alignment with organisational priorities, especially around governance. They also need to ensure that GenAI development practices occur in an environment that will ease the full deployment of these use cases across the business and generate a quick return on investment (ROI). Without such an assessment, the operator may encounter avoidable problems, including slow time to market and inability to comply with regulations. 

The adoption of a GenAI platform will create an environment that will enable operators to consistently manage the lifecycles of GenAI use cases; in the development, deployment and management of FMs; and the inclusion of these FMs within existing and new applications.  These platforms will ensure that operators adopt common development and management processes when dealing with the complexities associated with working with multiple FMs/LLMs. 

Analysys Mason’s recent research and insights highlight three strategies for GenAI platform implementation

Operators may decide to:

• build their own GenAI platforms based on proprietary FMs/LLMs they have created on their own
• build a GenAI platform based on commercially available or open-source FMs/LLMs
• partner with a third-party GenAI platform provider that offers a commercial off-the-shelf (COTS) GenAI platform solution.

However, selecting the optimal approach to implement a GenAI platform requires careful assessment of several factors: financial and technical resources are a significant consideration; operators will also require an approach agile enough to support their need to launch GenAI use cases quickly; and they should seek to minimise potential problems associated with vendor lock-in. Operators therefore need clear guidelines on how to select a GenAI platform implementation strategy.

To address this need, Analysys Mason researched the market and spoke to operators and vendors to determine the possible implementation strategies that operators can consider. We identified three strategies and analysed their benefits and requirements (Figure 1). These are covered in our recently published report, GenAI and the telecoms sector: three GenAI platform implementation strategies.