001✓ copiedspec
AI readiness teardown
A three-week fixed-fee assessment of your data and AI stack that ends in a decision, with a written verdict you keep whatever you decide next.
- Duration
- 3 weeks
- Commitment
- Roughly 8 days of my time, spread across three weeks
- Price
- Quoted per engagement
002✓ copieddetail
What is an AI readiness teardown?
An AI readiness teardown is a three-week fixed-fee assessment that tells you whether your data and AI programme can do what your board has been promised. It ends in a written verdict and a decision, and you keep the document whatever you decide to do next.
Most companies that call me have already spent real money. There is a warehouse, a team of fifteen to fifty, a cloud bill that grows every quarter, and a board that has started asking about AI. What is missing is a straight answer about which parts of that are working.
Why the assessment usually finds the problem one layer down
Most AI programmes fail one layer below where they get blamed.
The agent that hallucinates gets blamed on the model. The cause is usually that the knowledge it can reach was never curated, and nobody owns the corpus. The dashboard nobody trusts gets blamed on the BI team. The cause is usually two definitions of revenue and no named owner for either. I start at the bottom of the stack and work up, because every layer inherits the errors underneath it and multiplies them.
What actually happens in three weeks
Week one is listening. I talk to your people, and I ask to sit in the meetings where decisions get made rather than the meetings where data gets discussed. I want to see the moment somebody commits to a number, and find out whether anything your data organisation produced was in the room.
Week two is reading. The warehouse, the lineage, the cloud bill line by line, the model inventory, the last six months of incidents, the backlog, the org chart. Artefacts argue with the interviews more often than you would expect, and the gap between them is usually the finding.
Week three is writing. You get the assessment, the roadmap, and the stop-doing list. Then a two-hour readout where your leadership team can push back on all of it while I am still in the room.
What you get that a consultancy deliverable does not give you
A stop-doing list. Every assessment I have ever been handed as an operator told me what to start. None of them told me what to switch off, because that is the part that makes someone unhappy. The stop-doing list is usually where the money is.
What this costs
The teardown is a fixed fee, invoiced in two parts. You get the written assessment whether or not you do anything else with me. Roughly half of these convert into a longer engagement, and the half that do not are the ones where the honest answer was that you did not need me.
003✓ copieddeliverables
What you get
- Twelve to fifteen interviews across leadership, data, engineering, finance, and your loudest internal customer
- A read of the actual artefacts: the warehouse, the cloud bill, the model inventory, the incident history, the roadmap, the org chart
- A written assessment, twenty to thirty pages, that names what is wrong and why
- A twelve-month roadmap with sequencing and cost
- A prioritised list of what to stop doing
- One two-hour readout with your leadership team
Not for
- Companies with no data team yet. There is nothing to assess and I would be an expensive way to buy a dashboard.
- Anyone who needs a board deck by Friday. Three weeks is three weeks.
- Organisations that have already decided and want the assessment to agree.