Skip to content

Who is this framework for?

This framework is written for anyone accountable for how AI-assisted development happens inside a company. That includes:

Role What you'll use most
CTO, VP Engineering, Head of Platform Roadmap, Metrics taxonomy, Leadership charter worksheet
CIO, Chief Digital Officer Governance overview, Regulatory mapping, Vendor standards
CISO, Head of AppSec S.H.I.E.L.D., S.E.C.U.R.E., Incident response, Audit trails
Head of Developer Experience / DevEx Lead C.O.D.E.S., Metrics baselining worksheet, Pilot scope
Engineering Manager / Tech Lead V.E.R.I.F.Y., C.L.E.A.R., Pilot retrospective worksheet
Individual Contributor (developer, SRE, data engineer) S.C.A.F.F., R.E.F.A.C.T., Prompt library starter
Legal, Privacy, Compliance IP & licensing, Data handling, Regulatory mapping
Procurement Vendor selection scorecard
Internal Audit / Risk Risk classification, Audit trails
HR / L&D Training needs assessment
Board / Audit Committee Evidence overview, Regulatory mapping

Not sure where to start?

Pick the statement closest to your situation.

"We haven't decided whether to adopt AI-assisted development yet."

Read, in order:

  1. Philosophy: the framework's stance (augmentation, not replacement).
  2. Evidence overview: what the public research actually says.
  3. Leadership charter worksheet: decide, in one meeting, whether and under what conditions to adopt.

"Developers are already using Copilot/Cursor/Claude, we're catching up."

This is the most common starting point. Read:

  1. Phase 0: Readiness. Stabilise what exists before it grows.
  2. AUP drafting worksheet. Publish an interim policy within two weeks.
  3. Incident response. Have a plan before you need one.

"We've piloted and need to scale."

Read:

  1. Pilot retrospective worksheet: structured go/no-go.
  2. Phase 2: Scale: expansion playbook.
  3. Metrics taxonomy: what to measure to justify expansion.

"Legal/Privacy/Risk is asking hard questions."

Read:

  1. Regulatory mapping: crosswalk to SOC 2, ISO 42001, EU AI Act, HIPAA, PCI.
  2. IP, licensing & provenance.
  3. Data handling & residency.

"We're in a regulated industry."

Start with Public sector & regulated and then follow the "Legal is asking" path above.

Who this is not for (yet)

  • AI model trainers. This framework is about using coding assistants, not training them.
  • Pure-research teams. If your AI use is exploratory with no production path, the governance overhead here will be disproportionate. Skim Philosophy and return when something goes to production.
  • "No AI, full stop" organisations. You can still read Evidence overview as input to that decision, but the rest won't apply.

Changing roles?

Bookmark this page. The right starting point for a new CISO is not the right starting point for a new tech lead, and the framework is designed to be re-entered from any role.