Skip to main content

Framework Library

This is the canonical Framework section of the AI Playbook: every major framework used for AI consulting and end-to-end delivery, organised by job-to-be-done.

Each framework entry is written for practice, not posters: purpose, when / when not, numbered how-to, a filled Enterprise worked example (Apex Audit Partners), artefact, lifecycle stage, stage-gate contribution, failure modes and related frameworks.

Running client across the catalogue: Apex Audit Partners — mid-market financial auditing firm industrialising AI for engagement risk, journal anomaly detection, document extraction and working-paper drafting, under independence, EQCR and confidentiality constraints.

The navbar label Framework points here (not to a generic intro page).

How to use this library​

  1. Identify the decision you must make (ambition, problem, readiness, funding, design, trust, release, adoption).
  2. Open the matching category below and pick one primary framework for the workshop or gate.
  3. Read the Apex worked example as a facilitation pattern; adapt names and metrics to your client.
  4. Produce the stated artefact; attach evidence quality and an owner.
  5. Pass or fail the related stage gate before escalating investment.
  6. Use interactive canvases in the playbook app where available (MECE, Five Whys, SIPOC, VSM, JTBD, AI Canvas, RICE, SWOT, PESTLE, BMC, 7S, NIST AI RMF, RACI, RAID, ADKAR, business case, TCO/ROI).

Framework stack (end-to-end)​

Mobilisation → Strategy → Discovery → Readiness → Prioritisation
        → Commercial → Operating decisions → Architecture / data / ML
        → Responsible AI + Security gates → Delivery → Change
        → Operate (MLOps/LLMOps/AgentOps) → Scale or retire

No single framework covers the journey. Consulting frameworks decide what and why; technical frameworks decide how to build and run; governance and security frameworks decide whether work may proceed.

Categories​

CategoryUse when you need to…Page
Mobilisation & decisionsCharter the engagement and clarify who decidesMobilisation
StrategyDecide why AI, where to play, how to winStrategy
DiscoveryReplace assumptions with evidenceDiscovery
Readiness & maturityTest whether the organisation can deliver and sustainReadiness
PrioritisationRank and fund the portfolioPrioritisation
Commercial & valueProve economics and benefits ownershipCommercial
Architecture & engineeringDesign a coherent, operable systemArchitecture
Technical AI engineeringData, ML, MLOps/LLMOps, orchestration, domain toolkitsTechnical AI
Responsible AI & governanceMake trust, impact and evidence systematicGovernance
Security & privacyProtect users, data, models, tools and suppliersSecurity
DeliveryOrganise learning, build and releaseDelivery
Change & adoptionMake correct use stickChange

Methodology spine​

ArtefactRole
8D AI Solution Engineering FrameworkPrimary user journey: Define → Deliver
VALUE quality gateEvidence bar before escalation
Playbook page templateStandard page structure for new content

Companion long-form articles​

Selection cheat sheet​

DecisionStart here
Are sponsorship and decisions clear?Mobilisation — Charter, RACI, RAPID, RAID
Why invest and where?Strategy — cascade, Three Horizons, value-driver tree
What is really broken?Discovery — JTBD, SIPOC, VSM, Five Whys
Can we deliver?Readiness
Which use case first?Prioritisation — DVF, RICE, value–feasibility–risk
Does it pay?Commercial — TCO, unit economics, benefits plan
What system should we build?Architecture + Technical AI
Is it allowed and safe?Governance + Security
How do we ship?Delivery
Will people use it correctly?Change
Last updated on

Discussion

Comments​

Share feedback or questions about this page. No account required.

Loading comments…

Occasional notes on AI delivery.

Unsubscribe anytimePrivacy