Adopt AI faster, with control you can prove

The playbook

Train → Audit → Build

Safe AI adoption is a sequence, not a policy document. We work in three stages: people first, then visibility, then controls and scale.

1. Step 1 · Train

   Make your people safe users first

   Practical sessions on the tools your teams really use: what data can go in, how to check an answer before relying on it, and when to escalate.

   For

   Everyone using AI tools, and the managers who approve them.

   You get

   • Role-based training sessions
   • One-page acceptable-use guide
   • Checklist for reviewing AI output
2. Step 2 · Audit

   See where AI is used and where the risk sits

   We map every AI tool and use case across staff, software and vendors, then score each one for sensitive data, cost of errors, bias, security and human oversight.

   For

   Leadership, risk, IT and compliance.

   You get

   • AI use and risk inventory
   • Prioritised risk and readiness report
   • Quick wins and go / no-go list
3. Step 3 · Build

   Put the controls in place, then scale what works

   We set up the governance starter kit and the controls for each approved use case, add AI terms to vendor contracts, and build or configure the use cases with review points designed in.

   For

   Teams ready to launch or expand an AI use case.

   You get

   • Policy, approval route and roles
   • Controls and testing for each use case
   • Quarterly AI assurance review

Already trained your teams? Start with the Audit. Already know your risks? Go straight to Build.

02

What's included

1. 01/08

   AI use and exposure scan

   Interviews and a review of your software and vendors to build an inventory of AI tools and use cases, including unofficial "shadow AI", with the data, decisions and owner behind each one.

2. 02/08

   Risk and readiness assessment

   Each use case scored for business impact, sensitive data, cost of errors, bias, security, vendor dependence and human oversight, with the biggest gaps and quick wins first.

3. 03/08

   Governance starter kit

   A short acceptable-use policy, an approval route for new AI tools, clear roles and accountability, data-handling rules and an escalation process, sized to your business.

4. 04/08

   Use-case review and control design

   Before a launch, we define permitted uses, human review points, testing, record keeping and when to pause or roll back.

5. 05/08

   AI vendor and procurement review

   The right questions on data use and retention, security, subcontractors, limitations, incident notice and feature changes, built into your procurement and contracts.

6. 06/08

   AI audits and health checks

   Testing that agreed controls actually work: access, policy use, output review, change tracking and issue reporting. For in-house models we add data quality, performance, bias and drift.

7. 07/08

   Incident readiness and training

   A simple playbook for wrong or harmful outputs, data exposure, unexpected automated actions and vendor incidents, and training on the tools your teams really use.

8. 08/08

   Value and performance reviews

   Whether each AI use case saves time, improves quality or earns revenue once errors and rework are counted, with a clear recommendation to scale, change or stop.

Where you are in your AI journey

Related services