Already registered? Log in now to personalize your experience!
Description
Every AI coding agent runs on the same foundation models. The real differentiator is context: what information reaches the model, and how efficiently. In this speech we introduce Prism, an intelligence layer that sits beneath any coding agent, providing MCP-based code intelligence tools and a living architecture wiki. Benchmarks show 56% average token reduction and up to +12% higher solve rates on SWE-bench. This talk covers how Prism works, why context engineering matters more than model choice, and how teams can integrate it into existing agent workflows.
Swisper is a Swiss-built AI assistant that goes beyond chat to take action. It provides an agentic AI platform for banking and consumer applications, enabling teams to move from demo to production...
