Take the Free Scan
Claude Readiness Scan

How the Claude Readiness Scan works

Six questions, two axes, a deterministic score, four possible results. This page publishes the full model behind the scan, so you can judge the method before you trust the outcome.

What it measures

Readiness for Claude in an organisation is two separate questions that usually get collapsed into one. The first is about ground: can your environment carry a reasoning system? That depends on the quality of your systems of record, the integration surface available to reach them, and the maturity of your governance. The second is about intent: do you know, precisely, what Claude should achieve? That depends on whether you can name the task it improves, the metric that shows it, and the person who owns that number.

Projects stall on either axis independently. A team with immaculate data and no defined use case builds a clever demo that changes nothing. A team with a razor-sharp case and a weak record builds a system that gives articulate answers to yesterday's facts. A single readiness score hides which failure you are heading for; two axes name it.

The two axes

Foundation, the horizontal axis, measures the state of the ground: record quality, integration surface and governance maturity. Value clarity, the vertical axis, measures the sharpness of the intent: how specific and measurable the intended use case is.

How the scoring works

The scan asks six questions, three per axis. Each answer carries a fixed score, the answers on each axis are summed, and a threshold of five splits each axis into low and high. There is no weighting, no model and no judgement call behind the curtain: the same answers always produce the same quadrant. That is deliberate. An instrument built to start an honest conversation should not be a black box.

The four results

Sharp case,soft ground Ready to prove Early, honestly Strong ground,blurry target Foundation → Value clarity →

Ready to prove (strong foundation, clear value). The ground is solid and the case is sharp. The next move is a small, real proof: live data, real users, real permissions, measured against the target you already hold.

Sharp case, soft ground (clear value, weak foundation). You know what Claude should do; the record, access or governance it needs is not there yet. Fix the foundation first. Reasoning over a weak record produces confident, wrong answers, and trust spent that way does not refill.

Strong ground, blurry target (strong foundation, unclear value). The environment can carry Claude, but nobody has written down the number it should move. Sharpen the case before building anything: a use-case workshop costs an afternoon, a pilot without a metric costs a quarter.

Early, honestly (weak foundation, unclear value). Neither the ground nor the case is ready, and finding that out in three minutes is the cheapest version of that discovery. Explore deliberately: pick one candidate use case, define its metric, and test the record it would need.

The six questions

These are the exact questions the scan asks, in full, with every answer option. Nothing sits behind the start button.

  1. Q1 Foundation

    Where does the data your Claude use case needs actually live?

    • Mostly in one governed platform with clear ownership
    • In a few core systems, integrated in places
    • Spread across systems, spreadsheets and inboxes
    • Honestly, we are not sure yet
  2. Q2 Foundation

    If Claude needed to act on your systems tomorrow (read a record, update a case), what would that take?

    • Little: APIs and a permission model are ready
    • APIs exist, but permissions are unclear
    • Significant integration work first
    • Nothing is exposed today
  3. Q3 Foundation

    Who signs off AI usage, data handling and security in your organisation?

    • An established AI governance owner with a policy
    • Security and legal, engaged case by case
    • Under discussion, nothing formal yet
    • Nobody yet
  4. Q4 Value clarity

    How specific is your leading Claude use case?

    • One named process with an owner and a sponsor
    • A shortlist of credible candidates
    • Broad ambitions: productivity, a chatbot, "AI"
    • We are exploring what is possible
  5. Q5 Value clarity

    Could you measure success in numbers today?

    • Yes: baseline measured, target defined
    • We know what to measure, no baseline yet
    • We would know it when we saw it
    • Not discussed yet
  6. Q6 Value clarity

    If the task your use case covers goes wrong, what happens?

    • Errors are recoverable and human review is natural
    • Costly but fixable
    • We are in regulated or high-risk territory
    • We do not know yet

What the scan is not

It is a structured conversation-starter, not an audit. Six questions cannot see your architecture, your contracts or your politics, and a quadrant is a hypothesis to test in conversation, not a certificate. It is also not a lead trap: your result renders on the page the moment you finish, without an email address.

FAQ

How long does it take?
About three minutes. Six questions, multiple choice, no free text.
Do I need to give an email address to see my result?
No. The quadrant and its short read render on the page when you finish. Leaving an email gets you the fuller written read, and a senior consultant follows up personally. No SDR, no sequence.
What happens to my answers?
They are stored with your submission so any follow-up conversation is informed rather than generic, handled as set out in our privacy policy, with a six-month retention period.
Can the result be wrong?
Yes. A deterministic instrument is only as honest as the answers it is given, and organisations are usually harder on themselves in some areas than in others. Treat the quadrant as a starting hypothesis, not a verdict.