AI capability alone does not produce enterprise value. What produces value is an operating model that can absorb the capability into workflows where it is doing load-bearing work.
The AI Readiness Framework is five questions an enterprise should answer honestly before deploying AI into any business process. Each question is designed to expose a structural constraint that determines whether the AI investment will materialise as productivity or as expensive-but-ignored software.
Question 1 — Is the workflow this AI enters actually the right workflow?
AI deployed into a workflow designed for a different constraint set reproduces the workflow's original limits. If the workflow was designed for the throughput of a human team, an AI-assisted version of that same workflow will produce marginal gains rather than step-change ones. The reengineering answer is to redesign the workflow around AI-capable execution first, then deploy the AI into it. The technology-first answer is to bolt the AI onto the existing workflow and hope.
Question 2 — Where in the workflow is the AI doing load-bearing work, and where is it decorative?
The economic return of an AI deployment concentrates in a small number of steps where the AI genuinely replaces or augments human throughput. Elsewhere the AI is decorative — visible, tracked, adopted, but not moving the outcome. Enterprises that cannot distinguish the two before deployment usually cannot distinguish them after deployment either, which is why productivity data disappoints.
Question 3 — What are the reversibility constraints on this AI decision?
The AI capability landscape changes on quarterly time-scales. Irreversible commitments to a specific vendor, a specific model, or a specific deployment topology are structurally risky. The Design for Reversibility principle applies especially here. Prefer decisions that can be revisited without a fresh transformation programme.
Question 4 — What differentiating logic is at risk of drifting into the vendor?
When differentiating business logic gets encoded into prompts, fine-tunes, or workflow configurations inside a general-purpose AI vendor's product, that logic becomes available to competitors on the same commercial terms. The Protect Differentiating Logic principle applies. What is at stake is not the AI capability itself — that is commodity — but the operating model that surrounds it.
Question 5 — What P&L outcome is this AI investment actually being measured on?
Adoption metrics are lagging indicators of nothing that matters. Prompts served, users onboarded, and integrations completed do not produce P&L movement. The framework's final question is whether the AI investment is anchored to a specific business outcome — cost reduced, revenue captured, cycle time compressed — that is visible in the enterprise's actual results. If the answer is unclear, the investment is likely to disappoint regardless of the AI's technical quality.
What high AI readiness looks like
An enterprise with high AI readiness can answer all five questions specifically, not abstractly. The workflow has been examined. The load-bearing steps are named. Reversibility is deliberate. Differentiating logic is protected. The P&L outcome is measurable.
An enterprise with low AI readiness answers the questions with variants of "we'll figure that out during rollout". This is where the productivity gap opens and where six-months-later disappointment is generated.
Where AI readiness fits in the reengineering discipline
AI readiness is a specific application of the broader Digital Reengineering discipline. Both rest on the same sequencing claim: the operating model precedes the technology decision. AI happens to be the current era's technology decision that most acutely exposes this principle, because AI capability is advancing faster than most enterprises' operating models can absorb it.
Further reading
- Digital Reengineering: A Practical Guide for Enterprise Leaders — the discipline this framework applies to AI specifically.
- Why Most Digital Transformations Fail at the Org Chart, Not the Tech Stack — the failure pattern that repeats across every generation of enterprise technology, including AI.
- The Reengineering Playbook — the four-option decision framework that AI deployment decisions inherit from.
Sources
- Michael E. Porter — "What Is Strategy?", Harvard Business Review, November–December 1996. Anchors the differentiation-vs-operational-effectiveness distinction that underlies Question 4.