Yasmine Durand

Yasmine Durand

Miles Consulting Founder
#98325450 followers1 following
☀️ 2 day streak
>10,000All time
6 KP

About

I've been working in product for 15+ years, mostly on complex platforms, APIs and enterprise products. More recently, I've been spending a lot of time on AI and especially on what happens when AI stops just generating answers and starts taking actions in real systems. I'm the co-author of CROA, an open framework we're developing around that problem: how to keep execution authority under control while still letting agents do useful work. I also run Miles Consulting, where I work with companies on AI product strategy, getting AI initiatives into production, and agentic AI governance. I like building things, testing ideas in the real world, and talking to people who aren't afraid to challenge the assumptions behind them.

Badges

Tastemaker
Tastemaker
Gone streaking
Gone streaking

Forums

CROA p/croa•

5d ago

Where should execution authority live in an agentic system?

One of the assumptions behind CROA is that an AI agent should be able to propose actions, but should not be the final authority deciding whether those actions are allowed to change the state of a real system.

That sounds straightforward until you start looking at trajectories rather than individual actions.

Imagine an agent is allowed to expose up to 100 customer records. It performs three operations:

  • expose 40 records

  • expose another 40

  • expose another 40

•

2h ago

CROA - Constrain what AI agents can reach

CROA is an open, vendor-neutral architecture for governing what AI agents can actually execute. Instead of relying on the agent to respect constraints, CROA separates intelligence from authority and constrains execution against registered invariants before state change. The public Pilot makes the approach reproducible: run it, inspect the evidence, and try to break the claim.
View more