Field Notes
Notes from building Turnkeeper Ward.
Engineering insights, product updates, governance patterns, and practical guidance for building production conversational AI.
Latest dispatches
5 notes published
- Research
A Referral Is Not a Case Until Someone Can Defend the Link
Safety intelligence already treats a signal as a suggestion with provenance, not a verdict. The hard problem is what happens when that object has to meet a case file that does not share those rules.
- Company
Building a Shared Standard for Safety Intelligence
Turnkeeper’s goal is to help platforms exchange privacy-minimized safety signals, reconstruct evidence, and coordinate accountable human review without creating a global blacklist.
- Company
Protecting Children Means Controlling What Reaches Them
Historical product note: Ward now leads with privacy-minimized longitudinal case reconstruction for specialist review alongside existing safety systems.
- Research
The EU AI Act Is Turning AI Governance Into an Evidence Problem
The EU AI Act is moving governance from policy claims toward demonstrable evidence. For safety teams, the practical question is whether a decision can be reconstructed without turning evidence collection into surveillance.
- Research
The Five-Second Test for AI Governance
Can you prove which policy governed the AI action that just happened?
Case studies
Customer stories
Practical notes from teams exploring interoperable, evidence-first safety operations with Turnkeeper.
Case studies will appear here as design-partner stories are published.
Subscribe
Get Turnkeeper updates as we publish them.
Join the newsletter for product notes, field research, and operating perspective—sent when there is something worth reading.