All services
Agentic AI Security Workshop
Most agentic AI risk work starts from a framework nobody has run. This starts from agent fleets we operate in production, and ends with a model your engineers can build against.
£1,600
1-day package · £200/hour
1 day (8 hours) · starts with a 60-min call
Format
Design your safe-agentic-deployment model in one full day: guardrails, human-in-the-loop boundaries, agent identity and adversarial testing - the principles we proved running live agent fleets in production.
Who this is for
- Engineering and security teams about to put AI agents in front of real data or real customers
- Organisations that need an ISO 42001 or EU AI Act story before a customer or regulator asks for one
- Teams whose agents already work and who now have to explain why they are safe
- Security functions being asked to approve AI use cases without a model for doing so
What you leave with
- A guardrail and escalation model mapped to your own stack, not a generic diagram
- Agreed human-in-the-loop boundaries: what an agent may do alone and what it must hand back
- An ISO 42001 and EU AI Act alignment checklist against your actual use cases
- A written prioritised action plan within 48 hours of the day
Included
- 60-minute alignment call to plan the day and review your AI use cases
- Full-day facilitated working session with your engineering and security leads
- Guardrail + escalation model mapped to your stack
- ISO 42001 / EU AI Act alignment checklist
- Written prioritised action plan within 48 hours
Work with FeatherON
Bring in a team that ships
Book a focused hour on the problem you are actually facing - agentic AI risk, a compliance wall, a build that has stalled, or a board that needs answers. You will leave with a plan, not a pitch.