Govern AI with evidence, not spreadsheets alone.
Use RetailHawk to configure AI assessment and authorised scenario-testing programmes around the questions your governance team defines, with evidence, human review and controlled release.
Use published questionnaires as the assessment framework.
RetailHawk does not provide a native AI inventory or model-risk engine. AI systems can be represented as programme targets for a configured assurance programme, keeping the public claim bounded to the existing engine.
Governance questions
Capture ownership, purpose, intended use, oversight and other questions your governance team defines.
Supporting evidence
Attach supported documents, screenshots or images to the relevant answers.
Fixed versions
Publish the questionnaire used for the cycle so assessments remain attributable to a specific question set.
Observe behaviour the way a user experiences it.
Authorised testers can follow scripted journeys through systems you own or are authorised to test, recording outputs in structured answers and supported evidence. RetailHawk does not claim automated model evaluation or continuous monitoring.
Scenario scripts
Use the questionnaire to define prompts, expected checks and evidence fields.
Human observation
Testers record what occurred; no autonomous safety judgement is implied.
Cycle comparison
Run later cycles with the same question set where you need comparable reviewed results.
A customer-assistant scenario programme.
A synthetic programme can assess a customer-facing assistant against your own governance questions, run authorised test journeys, attach screenshots or documents and route each submission through human review before release.
1. Assess
Complete the published governance questionnaire for the configured target.
2. Test
Run the authorised scenarios and capture the observed response.
3. Review & release
A separate reviewer approves, returns or rejects before client visibility.
Keep standards and legal conclusions outside the software.
RetailHawk can organise evidence used in AI governance work. It is not a certification body and using it does not establish conformity with ISO/IEC 42001, the EU AI Act or any other law or standard.
No model certification
No automated safety, fairness or compliance certification claim.
No regulatory conclusion
Your organisation and advisers decide how the evidence maps to regulatory obligations.
UK hosting
Built and hosted in the UK.
Bring ownership, purpose, evidence and review dates into one assessment model.
| System | Owner | Purpose | Assessment focus | Review |
|---|---|---|---|---|
| Customer support assistant | Service | Answer routine queries | Human oversight · privacy · supplier controls | Quarterly |
| Demand forecast model | Operations | Stock planning | Data quality · governance · change control | Six-monthly |
Illustrative governance structure only. RetailHawk does not certify ISO 42001 compliance or guarantee compliance with AI legislation.