Design partner story · Ride-hailing
Letting a large engineering team use AI agents without putting rider and driver data in reach.
A live design partnership focused on real developer workflows, real enterprise systems and measurable control.
IndustryRide-hailing and mobility
Agents in useClaude Code and Cursor CLI
StatusDesign partnership active
The challenge
The risk arrived before the control.
Engineering teams had adopted AI coding agents quickly and bottom-up. Agents were running with each engineer’s own access, which can include production systems, rider and driver data and payment services. The security team had no inventory of which agents ran where, and no way to limit what they could reach without slowing engineers down.
What they’re doing with RayShield
- Discovering AI tools and agents across a pilot group of engineering machines
- Mapping the blast radius of existing agent setups by team
- Wrapping coding agents in Aegis so only approved connections, repositories and connectors are reachable
- Keeping credentials away from agents with runtime-only secrets
What success looks like
- Complete inventory of AI agents across the pilot fleet
- No agent able to reach production data outside its policy
- Engineers keep the same workflow, with no new friction