Simulated User

concept · updated Jul 30, 2026

person concept tool org talk claim — click a node to jump to its page; hover an arrow for the relation

A Simulated User is a synthetic stand-in for a human participant in an agent interaction, implemented as an LLM driven by its own prompt rather than a live human providing input.

How It Works

Rustem Feyzkhanov (Snorkel AI) describes the pattern directly: rather than requiring a real user to interact with an agent during testing or evaluation, "you can simulate the user. In this case, that becomes effectively LLM with its own prompt." (9:21) This makes the simulated user a peer component within the broader agent system — an LLM-backed role that generates inputs, responses, or queries in place of a human counterpart.

Role in Agent Simulations

Feyzkhanov advocates for this technique in the context of moving from agent traces to full agent simulations. By replacing the human side of a conversation with a prompted LLM, it becomes possible to run end-to-end agent interactions autonomously — enabling scalable testing, evals, and data generation without human-in-the-loop overhead. (9:21)

Points of Disagreement

No dissenting views on the Simulated User concept appear in the current material.