Research

AI model evaluation research focused on testing the reliability, task completion accuracy, and safety of AI customer service agents across multi-step workflows.

Research

Best LLM for an AI customer service agent: we tested 12 models on real support chats

We gave 12 models the same prompt, the same 25 tools and the same messy customers, then graded 7,218 runs. Some models we expected to win didn’t, and the most expensive one didn’t come first.

Cevanos ResearchOctober 27, 202612 min read

Read

Start with answers. Add more automation when you’re ready.

Build your first agent, then explore the controls and testing tools as your needs grow.

Start free Explore the platform