RSECon25: Beyond Traditional Apps with LLMs

My accepted talk for RSECon25 on how Large Language Models might replace traditional app interfaces through conversational agents

My talk just got accepted for RSECon25¹ in September! I’m going to share how Large Language Models (LLMs) might just ditch the need for traditional “apps” and programs.

Most software still talks in buttons, charts, forms, pop-ups. But what if you could skip them entirely and get your data or perform actions the way you talk to a person?

In the walkthrough, I’ll show how LLMs (like ChatGPT) can become your front-end, powered by a new thing called the Model Context Protocol (MCP)². It lets you plug in real-time, structured data directly into a conversational agent.

I’ll demo how to build an MCP server, show what works (and what doesn’t), and share patterns I’ve seen when replacing traditional apps with this approach.

🗓 9 to 11 September 2025
📍 University of Warwick
🎤 RSECon25 – Ninth Annual Conference for Research Software Engineering

References

1.

RSECon25 – ninth annual conference for research software engineering. (Society of Research Software Engineering, 2025) (Conference details).

2.

Model Context Protocol. (Anthropic) (Protocol for connecting LLMs to structured data sources).