Beyond LLMs: Decision Models for Smarter AI Agents
Most steps in an agent pick from a short list instead of writing prose. Decision models such as Jev and the OpenAI Decisions API return typed answers with calibrated probabilities, so code can branch on them. The live demo is a booth follow-up agent on AWS. Step Functions runs the workflow, a Strands agent writes the email, Jev decides before and after it, and code owns the threshold.