Monitor · In development
Ongoing measurement of how your AI treats people.
Your live conversations, measured on a continuing basis against the same standard as a Diagnostic Record, so a change in how your system treats people is visible as it happens.
Why continuous
Systems change. The record shows when.
Releases move behavior
A new prompt, model or guardrail can shift how a system treats people. Monitor is being built to show when that happens.
Drift is a pattern
No single reply has to break a rule for a conversation to go somewhere it should not. We measure where it goes.
A versioned yardstick
Every result names the version of the standard it was measured against, so a change in the system can be told apart from a change in the measure.
How to start
The first step is a Diagnostic Record.
- Diagnostic Record · Open for pilots
Measure now
A Diagnostic Record gives you a first read of how your system treats people.
- Monitor · In development
Monitor next
When Monitor opens, pilots start with the teams who already have a Diagnostic Record.
Monitor is not for sale yet, and we take no pre-orders.
Monitor updates.
Ask to hear when Monitor opens. If you run AI that talks with people, a Diagnostic Record pilot can start now.