Designing systems that stop breaking people.
We design the controls that let you stop, reverse, and correct a high-stakes system: who can halt it, how a bad decision gets undone, and how the people it affects get recourse.
For health plans, healthcare AI companies, and investors who have to show that a system can be stopped, reversed, and governed. The method is open: ethotechnics.org.
How it works
Every engagement does three things: maps where the burden lands, rehearses the failure, and names who decides.
Map the burden
Find where effort, confusion, and delay pile up in a workflow, and who absorbs them.
Rehearse the failure
Drill escalation, override, and rollback with the people who will run them.
Name the owners
Write down who decides, who escalates, and how review happens, so decisions hold under pressure.
How an engagement works
- 1
Map one workflow. Show where burden lands and who holds stop authority.
- 2
Rehearse failure scenarios. Practice escalation, override, and rollback.
- 3
Deploy with the instrumentation, escalation paths, and maintenance routines the team will keep running.
Services
Fixed-scope engagements, priced by format.
- Clinical AI safety evaluation: $35–50K.
- VC diligence assessment: $15–25K.
- Governance readiness sprint: $10–15K.
Advisory
One-hour safeguards review
A 60-minute session with the founder. We check where your system's boundaries are, how quickly it can be stopped and reversed, and leave you a written list of fixes.
$1,500 per session
What the review covers
- Boundaries and clocks, checked against framework standard STD-01
- Contestability, fallback, and escalation paths, checked against STD-02
- The failure modes that would do the most harm, mapped
- A written roadmap of fixes, with the validators to run
If this sounds like your system, get in touch.
Before you book
Answer three questions, then pick a time. Your answers travel with the booking.