AgentOps

Our first thesis focused on MSMEs. Through customer research, we identified that the problem we were exploring did not have sufficient willingness to pay among potential customers. This led us to invalidate the idea and, more importantly, taught us to prioritise understanding and validating the problem before becoming attached to a solution. We then explored the EdTech space. After reaching out to potential customers and validating the problem, we found that the workload involved was not significant enough to create a strong need for the proposed workaround. Since customers could reasonably manage the task themselves, we concluded that the problem did not represent a sufficiently strong pain point and invalidated the thesis. Our third exploration was in real estate. Customer research helped us understand that the issue we had identified was primarily a matter of convenience rather than a genuine, recurring pain point. This distinction led us to invalidate the idea rather than force a solution onto a weak problem. Finally, we explored the finance domain, where we researched the issue of garage invoice inflation. We investigated the problem, its underlying causes, the stakeholders involved, and the potential impact of inflated invoices on customers and the broader ecosystem. Across these explorations, the biggest learning was to remain research-oriented rather than solution-oriented. Instead of trying to make an idea work, we learned to actively look for evidence that could validate or invalidate our assumptions. The internship taught us that identifying a real and meaningful problem is often more valuable than prematurely building a solution. The team explored AI-driven agentic frameworks across multiple verticals - Finance, Real Estate, EdTech, and MSME - applying structured problem discovery rather than starting with a solution. Verticals explored: - Finance: AI agents for retail investment research, portfolio monitoring, and regulatory compliance summaries for MSMEs. - Real Estate: Agentic systems for property discovery, document verification, and rental market analysis. - EdTech: Personalized learning path agents that adapt to student performance in real time. - MSME: Automated vendor discovery, invoice reconciliation, and cash flow forecasting agents for small businesses. Key learnings: Understanding the problem is the first step towards creating meaningful change. The team developed skills in market research, stakeholder interviews, feasibility analysis, and rapid ideation across domains.

The internship gave us the opportunity to look beyond predefined problems and explore ideas from the ground up. It taught us how to approach unfamiliar domains, question existing assumptions, and use research to identify meaningful opportunities for innovation.