AI Readiness
Assessment.
A gate on an AI programme, before the commitment rather than after it. Five pillars, one answer, and the reasoning written down.
CTO · CIO · CRO · Transformation lead
When it applies
- A pilot is about to be commissioned and nothing defines what it must prove.
- A pilot worked, and it is about to justify a production commitment.
- A system is close to going live with no answer to who is accountable when it errs.
The five pillars
Assessed independently, reported together
Strategic Readiness
What the programme is for, who owns the outcome, and what it has to prove before it earns a production commitment.
Data and Process Readiness
Whether the data exists, is good enough, and reaches the system in the shape the use case assumes.
Technology and Architecture Readiness
Whether what is designed can be built, operated and changed at the scale being promised.
Governance and Risk
Who is accountable when the system errs, where the data actually goes, and what has to be true before it touches a customer.
Execution Readiness
Whether the organisation can run it after go-live: the operating model, the roles, the escalation path.
Most reviews stop at technical feasibility. Governance and Risk is the pillar that decides whether the thing survives contact with production, and it is the one most assessments leave out. It is also where the answers are least likely to be what the organisation assumes: a policy adopted from someone else's regulation with no mechanism to enforce it, or a deployment in a local data centre whose queries are answered in another region.
What it closes on
One of three answers, never a maybe
Go
The programme is ready. Proceed on the terms defined.
Conditional Go
Proceed once the named conditions are met. The conditions are specific and testable.
No-Go
Not ready. The reasons are documented, and so is what would change the answer.
An assessment that cannot return a No-Go is not an assessment. That is the point of bringing in someone who is not building it.
What we do
We work from evidence rather than opinion: your systems, your documents, and the people who own the outcome. Where it helps, we prove a pillar on a sample of your live data before the rest is committed.
What it produces
A per-pillar read, the conditions attached to each, and one closing answer the programme can be held to.
What it isn't
Not the readiness review written by whoever is building the system. The independence is the product.
On the assessment, Raphaël Peyret covers use case selection, stakeholder alignment, governance, and the non-technical risks that keep working AI out of production.
What you are committing to
Typical shape, not a fixed package
How long
Typically three to four weeks where one use case is in scope. Longer where the programme spans several and the stakeholder count climbs.
Your time
The programme owner, and access to the stakeholders who hold each pillar.
What you keep
A per-pillar read, the conditions attached to each, and one closing answer.
Three things move the length: how many of the five pillars sit in phase one, how many stakeholders have to be interviewed, and whether we prove a pillar on a sample of your live data before the rest is committed. Organisational complexity is the largest factor: a single programme with one owner and a multi-entity transformation with four are different pieces of work. Most of the calendar is scheduling and evidence gathering rather than working time. We scope it against your decision date, not a fixed package.
On the record
“… He doesn't jump to solutions. He gets to the crux of the problem first, builds solid hypotheses, and does the foundational research to back them up. The rigor was real. …”
Recognise your programme in this?
That is enough to start. No business case, no budget line, and no answer yet to what it has to prove.
Leave an email and we will come back to you.
No obligation either way. We will come back with a tailored answer, or tell you it is not us.
Or write to us directly at contact@sha-rp.com.
Also for enterprise