Built a tool for testing AI agents in multi-turn conversations
We built ArkSim which help simulate multi-turn conversations between agents and synthetic users to see how it behaves across longer interactions. This can help find issues like: – Agents losing context during longer interactions – Unexpected conversati…