TM Forum Innovation Hub: Generative AI pilots simplify knowledge sharing and ODA use cases

TM Forum launched Innovation Hub with the ambitious goal of addressing the “telecom industry’s most critical challenges” and to “reignite growth” by providing practical solutions that are immediately usable in operations. Google Cloud has been a founding member and an active player in two of the three Innovation Hub projects from its pre-pilot days. We are delighted to share our perspective on the substantive progress that has been made by Accenture, Deutsche Telekom (DT), Google, Jio, Orange, Telenor & Vodafone in the pilot phase, demonstrating the value of collaboration-at-speed to apply emerging technologies such as Generative AI, to address the industry’s significant challenges.

As the Innovation Hub was getting off the ground to develop new solutions, George Glass, TM Forum CTO, pointed out an interesting conundrum: identifying and leveraging even the existing TM Forum assets is not easy today! TM Forum has extensive assets for addressing many critical challenges faced by the industry, such as autonomous networks, sustainability or cloudification. Glass challenged us to make the existing assets more easily usable using Generative AI. In addition to enabling Communication Service Providers (CSPs) to benefit more easily from the TM Forum assets, the learnings from the effort could also help them to better glean insights from their own corporate assets while addressing security, privacy, accuracy, performance and scalability needs. This very practical requirement resulted in one of the first Innovation Hub projects leveraging Gen AI – TM Forum Guru!

Everyone needs a guru

Two TM Forum engineers with input from Google Cloud built a bot called the TM Forum Guru, ingesting about 5000 web pages and documents. This bot, built in less than two weeks, was able to answer some of the questions suggested by the TM Forum experts while showing it is possible to avoid some of the common concerns about Generative AI such as hallucinations. After this initial validation, a small team from Accenture, DT, Google, Jio and Vodafone has been adding access management, multimodal interactions, multi-turn conversations, code generation from Open Digital Architecture (ODA) specifications and other capabilities. The team is eagerly looking forward to showcasing the current TM Forum Guru at DTW Ignite ‘24! During the same event, CSPs will have a hands-on opportunity to put TM Forum Guru to use to solve some of their real problems in a hackathon!

Plug and play - AI, Cloud Native ODA

ODA enables CSPs to create new and differentiated services that improve productivity, reduce maintenance and integration costs, whilst enhancing customer experience. Another Innovation Hub project Google Cloud has been active in is the ODA Accelerator, whose goal is to make it easier for CSPs as well as technology vendors to adopt ODA while making the ODA runtime environments scalable, secure and responsive to rapidly changing business needs by leveraging cloud and AI.

Google developed an ODA Canvas on Google Cloud, and working with Matrixx, Ericsson and Orange, was also able to demonstrate ease of deployment, management and interoperability between components from both of these companies. Since then, Google Cloud has been actively working with DT, Orange, Vodafone and other partners to address a number of challenges from accommodating legacy systems to running AI/ML-enabled ODA components, able to support use cases such as fraud detection and AIOps.

We are excited to enable CSPs to leverage ODA to innovate faster, without having to wait to retire their legacy systems, by using Gen AI to generate connectors, managed using API management solutions.

Let us close with some key lessons we have learned during the pilot phase:

1. Embrace diversity of thought

The two projects have brought together technology and business thinkers representing a cross section of the industry including CSPs, technology providers and system integrators. Collaboration across this diverse group was instrumental in identifying significant industry challenges and promising technologies to address them.

2. Keep pace with change

Technologies such as Gen AI evolve rapidly, making it important to iterate swiftly. For example, we recently switched to newer LLMs with significantly larger context windows and multimodal reasoning capabilities. We had to test the impacts of such changes. Our Gen AI platform features such as side-by-side model comparison made such inevitable changes easier.

3. Make it real to the organization

Innovation Hub participants such as DT and Accenture have drawn on the opportunities and the challenges they encounter when deploying Gen AI in their own enterprises to shape Innovation Hub projects while simultaneously applying Innovation Hub learnings within their companies – thus ensuring that the innovations are immediately practical.

We look forward to sharing the outcomes from the TM Forum Innovation Hub pilot phase at DTW Ignite ‘24! We invite you to visit the dedicated kiosks at the DTW TM Forum Industry Showcase and experience the innovation first hand!

The author thanks Brian Kracik, Jen Hawes-Hewitt, and Priya Saxena for their valuable input.