The AI Readiness Assessment Framework: 6 Weeks, 5 Vectors, a Written 90-Day Plan

Most AI readiness assessments produce 60-slide decks that nobody reads. A useful one produces a 40-page document with a prioritized use-case portfolio, a risk register, and a 90-day action plan that a single named owner can execute on Monday. Here is the framework we use.

Quick answer

Our AI Readiness Assessment runs for six weeks, interviews 10–15 stakeholders, audits five organizational vectors, and produces three deliverables: a written assessment, a prioritized use-case portfolio with ROI estimates, and a 90-day action plan. It is designed specifically to prevent the failure patterns we covered in Why 95% of Enterprise AI Pilots Fail.

The five vectors

1. Data infrastructure

Questions answered: Where does your data live? How clean is it? Can it feed an AI system in production without months of re-engineering?

We audit: source systems and pipelines, data quality and lineage, embedding and vector-store readiness for RAG, classification and access controls, and real-time-versus-batch availability. The output is a data-debt register — what must be fixed before any AI system can ship against this data.

2. Talent

Questions answered: Do you have the skills in-house, and if not, what is the shortest path to acquiring them?

We map roles to capability: data engineering, ML/LLM engineering, prompt engineering, evaluation, AI product management, and change management. For each gap we propose one of three paths: hire, train, or partner. Training is usually the right call — our AI Academy exists precisely for this reason.

3. Workflows

Questions answered: Which workflows in your organization are high-leverage AI candidates, and how would they actually change?

We shadow or interview the teams running the candidate workflows. For each, we produce a before/after diagram: what the person does today, what they would do with AI, and what the measurable metric would move. Workflows that cannot produce a measurable metric are deprioritized — they are the source of most pilot failures.

4. Governance

Questions answered: What must you put in place before legal, security, and compliance can sign off on AI in production?

We audit: existing policies around data privacy, vendor risk, acceptable use, and audit logging; gaps against regional regulation (for Lebanon-based clients, Law 81/2018; for multinationals, GDPR and the EU AI Act); readiness for the 3-Tier Safety System; and incident-response plans.

5. Competitive context

Questions answered: What are your competitors, peers, and upstream/downstream partners doing with AI, and what does that mean for your timing?

We scan: recent AI-related announcements from direct competitors, regional AI initiatives in your industry, emerging vendor ecosystems, and partner/supplier AI capabilities that could change your cost structure. This informs sequencing — some use cases are urgent because a competitor is moving, others can wait.

The six-week sequence

Week 1 — Scoping & kickoff

• Executive sponsor alignment on goals and success criteria for the assessment itself.
• Stakeholder roster: 10–15 people across functions.
• Data room setup (NDA-protected, access-logged).

Weeks 2–3 — Vector deep-dive

• Interviews with each stakeholder (45–60 min, semi-structured).
• Data infrastructure audit (read access to metadata, sample queries).
• Workflow shadowing for the top three candidate use cases.
• Policy and governance review.

Week 4 — Synthesis

• Use-case portfolio with ROI estimates and prerequisite dependencies.
• Risk register with severity, likelihood, and mitigation owner.
• Capability-gap matrix.

Week 5 — Drafting & internal review

• Full document draft.
• Two rounds of internal QA review inside our team.
• One round of client technical-lead review for factual accuracy.

Week 6 — Executive readout & handoff

• 90-minute readout to the executive sponsor and nominated delivery owner.
• Q&A, document revisions, and handoff of the 90-day action plan to a named owner.

What you walk away with

Three concrete artefacts:

1. A written assessment document (25–40 pages). Executive summary, vector-by-vector analysis, use-case portfolio with ROI estimates, risk register, 90-day action plan, 12-month roadmap.
2. A prioritized use-case portfolio. Typically 8–15 candidates, ranked on three axes: expected value, feasibility, and strategic fit. Each candidate lists the data dependencies, talent requirements, and governance prerequisites.
3. A 90-day action plan with a named owner. The sequence of moves to make in the next quarter to unblock the top-priority use case — not the entire roadmap.

The 90-day plan is the single most important output. Ninety days is short enough to preserve urgency and long enough to move a meaningful needle. A named owner makes it accountable. Without both, assessments end up on a shelf.

Next step

If you think your organization is ready to start deploying AI but want an outside view on readiness first, the AI Readiness Assessment is our default opening engagement. It is the fastest path from “we should be doing something” to a concrete plan a team can start executing on Monday.