Use Case
Our team participated in a 2 days Hackathon at DTW24 Ignite in Copenhagen with the challenge of creating a business use case and implementing a demo using TMF AIVA, Google Cloud Platform (GCP), and internal enterprise data. TMF AIVA, powered by Google’s Gemini, is an enterprise search agent that provides access to TM Forum documents, standards, and best practices. We leveraged GCP’s capabilities to deploy a solution based on Generative AI.
​
Team members (from right to left):
Balint Kardos: UX/UI Product Designer
Mark Harkay: UX/UI Product Designer
Birgit Kruckewitt: Product Owner
Sergei Lukin: Architect
Rishat Alkhamov: Solution Designer​​​
Problem
Product Managers in our organization often face the challenge of identifying the right TMF Open API specifications to address specific business needs. Manually analyzing all 70 TMF Open APIs to find the relevant ones is time-consuming and inefficient. Additionally, aligning existing APIs with TMF standards and determining whether to reuse or develop new APIs can be a complex task, leading to potential redundancies and misalignments within our IT landscape.
Solution
We proposed a scenario where a Product Manager describes their problem in natural language, such as wanting to implement a new marketing program for private users. This statement is sent to TMF AIVA, which then provides a list of relevant TMF Open API specifications. For instance, TMF AIVA might recommend TMF671 Promotion Management, TMF629 Customer Management, and TMF622 Product Order Management.
​
The second step of our solution involves using Vertex Agent (vector index/search) to search for similar APIs within our internal DT API Catalog. This allows us to identify which existing APIs align with the TMF standards provided by AIVA. Based on this analysis, we can decide whether to reuse an existing API or develop a new one, ultimately reducing the number of APIs in use and improving their alignment with TMF standards.​​
My role
I designed a UI that allowed the Product Manager to view the TMF Open APIs recommended by TMF AIVA, along with their matching internal APIs. The APIs were listed in a table format, displaying their Name, Description, and Relevance, with the highest relevance highlighted to indicate the best match for the PM. This design ensured that the PM could quickly and effectively identify the most suitable APIs for their needs.
​
I also played a key role in preparing the presentation of our use case to the jury, ensuring that the highly technical aspects were communicated in a clear and understandable manner. My contributions included simplifying complex information and making it accessible to a broader audience.
​
You can find our presentation, along with those of our competitors, on the TM Forum Wiki page.​​
​
Even though we just missed out on first place, we created a solution that holds great potential for future use. Beyond the hard work, we had a lot of fun, and Copenhagen is an incredible city.​​
