TM Forum and AWS Empower CSPs in the AI Era

“The pace of Generative AI innovation and experimentation in the telco sector has been impressive. In less than a year, we saw tangible use cases emerge that CSPs and telecom operators now seek to deploy at scale. Helping the industry to achieve their ambitions and create measurable outcomes by AWS working with TM Forum on supporting generative AI is a great opportunity for us,” said Ishwar Parulkar, Chief Technologist, Telecom at AWS.

Drawing on anonymized, aggregated data from in-depth surveys conducted with more than 200 AI decision-makers among CSPs worldwide, GAMIT enables CSPs to benchmark their generative AI (GenAI) maturity against regional peers and global leaders, identify priority use cases and rapidly move from development to production at scale.

George Glass, Chief Technology Officer of TM Forum, commented, “The insights gained from GAMIT will inform the future direction of TM Forum’s collaborative initiatives, including our Catalyst projects and Innovation Hub, ensuring that members have the data, resources and support they need to thrive in an AI-driven future.”

GenAI Struggles Persist Among Telcos

The AI opportunity for telcos is considerable as the initial findings from the data used to build GAMIT show that the adoption of GenAI across the industry remains in its early stages. Many CSPs are still struggling to move from bolt-on solutions to integrating AI as a built-in, native element across all operations.

Data from GAMIT reveals that that only 25% of operators feel equipped to leverage advanced techniques like fine-tuning, Retrieval Augmented Generation (RAG), and prompt engineering for accurate results, with just 16% confident in using these methods to optimize costs and ROI.

Leadership roles focused on AI are limited, with only a third of CSPs having a CXO dedicated to AI strategy. Furthermore, only 14% of operators have more than 10 GenAI use cases in production, and just 33% report having a reliable GenAI pipeline. The most common barrier to advancing GenAI initiatives is the lack of accuracy in proof-of-concept stages.

Six Pillars of GenAI Maturity

Underscoring the need for wider industry collaboration on, and understanding of, the challenges CSPs face from implementing GenAI at scale, these findings have inspired the benchmarking model that underpins GAMIT to be based around six pillars of GenAI maturity:

Technology Understanding and Maturity: Evaluating and adapting foundation models to address industry specific use cases to deliver swift return on investment (ROI).
Organization, Responsibilities, and Skills: Embedding the necessary skills and roles across the organization to prioritize and execute AI deployments consistently at pace.
Data Readiness and Availability: Ensuring fast and easy access to high-quality data from the telecom domain across the complete CSP organization to monetize data assets successfully and to enable fine-tuning models and model performance at inference.
Governance, Privacy, Compliance, and Security: Integrating guardrails and policies to enable privacy and security while maintaining velocity.
Business Objectives: Enabling CSPs to quickly apply GenAI to high value workflows to realize a step-function in value.
Taking AI Use Cases into Production: Accelerating from proof of concept to live deployments with cloud native composable building blocks and design primitives.