Measure · Open for pilots
An independent Diagnostic Record of how your AI treats people.
A Diagnostic Record is an independent assessment of your system against a named version of Ikwe's standard, the EQ Safety Benchmark: a score for every turn and one for the whole conversation, the path each conversation took, and the turns where it fell short, with a written report. Nothing is installed in your system.
A Diagnostic Record and a written report.
A test endpoint, an access key and a signed scope. We do not need your codebase, your system prompt or your users' identities.
Pilots start with a free sample read of your system. A full Diagnostic Record is priced on scope.
A Diagnostic Record comes back in under a week.
Two ways in
Pick what you can share. Both are scored the same way.
Scenario run
We run your system through our designed scenarios against a test endpoint you give us. No user data leaves your hands. Best for a first read.
Open for pilotsHistory review
We score a stretch of your real conversation history. Offered once a data agreement and a privacy review are in place.
By agreementWhat happens
What a pilot looks like.
- 1
Agree the scope
Which system, which version, which conversations. You sign first.
- 2
We run it
Every AI reply is read with everything said before it and scored against the standard.
- 3
A reviewer checks it
A reviewer on our team checks every Diagnostic Record before it is issued.
- 4
You get the record
Dated, stamped with the version of the standard, and showing the path each conversation took. Yours to hand on, or not.
Diagnostic Record · illustrative
What comes back
A Diagnostic Record and a report.
A score for every reply
Eight behaviors, combined into one score from 0 to 100 for each turn, with any Safety Gate flags beside it, and one score for the whole conversation.
The turns that matter
Every finding traces to its turn, with the transcript around it.
A plain-language finding
What happened in the conversation, written for the people who will read it.
Bands describe where a score falls on the current version of the EQ Safety Benchmark. Clinical validation of the benchmark is under way, and pilot results are reported as exploratory. A Diagnostic Record is a diagnostic, not a certificate.
Start with a pilot.
Pilots are for companies running AI that talks with people. Send a note about your system, and we will agree the scope in writing before anything connects. A Diagnostic Record is also where Monitor starts.