When we were given the use cases for the LBS AI Agent Cup, the Executive Education challenge immediately stood out to me. I could draw a direct line to my time in investment banking. At ING, we would receive around ten RFPs every week, and the way we responded to them was painfully manual and time intensive. Reading about Executive Education teams spending 10 to 14 days doing essentially the same thing, I knew this was a problem I understood well enough to build for.
Together with my teammate Praneeth, we built Groot. The system uses a multi-agent architecture. An extraction agent parses the incoming RFP document to identify requirements and evaluation criteria. A matching agent searches historical proposals through vector search to surface relevant past responses and case studies. A generation agent then drafts a tailored response by combining what it extracted from the RFP with the best matching historical content.
What started as a hackathon project has now turned into something real. I continued working on the project and have been demoing the project to senior leadership, piloting the tool within executive education.
