Service
AI Threat Modelling
Who it is for
Teams designing an AI feature or agent who need design decisions recorded before build, and buyers who ask what was considered.
When clients call
- An agent or retrieval pipeline is in design and the abuse paths are unwritten
- A security review has been requested and no threat model exists
- A model, tool or provider change alters what the system can do
Scope
- Asset, actor and trust-boundary map for the AI system in scope
- Abuse cases and attack paths drawn against OWASP LLM Top 10, OWASP Agentic AI threats and MITRE ATLAS
- Data-flow review across prompts, retrieval, tools, logs and model providers
- Control decisions, residual risk and the owner accountable for each
Method and standards
- OWASP Top 10 for LLM Applications 2025
- OWASP Agentic AI threats
- MITRE ATLAS
- NIST AI RMF 1.0
Deliverables
- Threat model for the system in scope
- Abuse-case and attack-path register
- Control decision record with owners
- Prioritised remediation list
Engagement arrangements
Dates, responsibilities and any ongoing review cadence are agreed in the written scope. Retesting is included where stated in that scope.
What happens after the scoping call
Written scope and fixed-fee proposal within one business day of your scoping call. The scope names the standards applied, the deliverables, the dates and the people doing the work.
Questions
How is this different from the application security assessment?+
Threat modelling is a design activity: it establishes what the system must defend and which control closes each path. The application security assessment then tests whether those controls hold in the running system.
Do you need production access?+
No. Threat modelling works from architecture, data flows and intended permissions. Where a running system exists, its configuration is reviewed as evidence.
How is the fee set?+
The written proposal defines the fixed fee, responsibilities and deliverables. Any change to scope is agreed in writing before additional work begins. Ongoing advisory includes an agreed review cadence and availability.
How is our evidence handled?+
Evidence is held under NDA, encrypted at rest and in transit, restricted to the named engagement team and destroyed on the schedule stated in the scope. It is never used to train any model.
Tell us what changed.
Thirty minutes, no obligation. Written scope and fixed-fee proposal within one business day of your scoping call.
Book a Strategic AI Readiness Assessment