Put the problem on paper first.
Three weeks, one industrial AI decision, one fixed price. We leave your team a sponsor-signed baseline and a written acceptance test they can run without us — and $2,500 of our own fee is only invoiced when the scoped pilot passes that test. We sell no implementation, so the diagnosis is the whole product.
the problem it solves
Most AI initiatives fail before a line of code is written — they fail at the problem statement. The number-one documented cause of AI project failure is misunderstanding what the project was for; the common outcome is a pilot with no measured impact.
Every readiness audit on the market answers this with a forward-looking opinion the seller carries no exposure on — usually sold as a down payment on the seller's own implementation work. The party that diagnoses your AI problem usually sells you the cure. This audit is the opposite instrument: it ships a testable artifact, from a firm charter-barred from selling the build, with part of its fee riding on whether the pilot ever passes.
what it is
One AI use case, converted in three weeks into three artifacts your own team can act on and check — a sponsor-signed baseline metric, a written vendor acceptance test on your own data, and a scoped-pilot definition — plus a Go/No-Go decision on the record. Not a maturity model, not a roadmap: those are opinions.
One decision, one price.
$12,500 flat for one AI use case. No hourly billing, no scope creep, no change orders inside the three weeks. A second initiative is a second audit, at the published price.
A number the sponsor signs.
The workflow restated as metric, current value, target value, owner — measured system-of-record first, each figure tagged [counted] or [estimated], and signed by the person who owns the P&L it touches.
A test your team runs alone.
A vendor acceptance test with measurable thresholds, a sampling rule, and a binary pass/fail — executable on your own data with no Tektari involvement, usable verbatim as a contract exhibit against any vendor.
$2,500 rides on the result.
The final tranche is invoiced only when the scoped pilot passes that acceptance test — never if it fails, never if the pilot is never run. A "don't run this" finding is a fee-costing recommendation for us.
the three-week plan
One use case on the record, a baseline the sponsor signs before any threshold is negotiated, and an acceptance test your team dry-runs until they can execute it without us. If a step can't survive the honest version, that finding is the audit's value — it gets said, not papered over.
- Week 1
One use case, on the record
Kickoff with the sponsor and the initiative owner; if you bring several, we triage to one — the audit covers one use case by design. Short interviews with the person who does the work and the person who signs the metric. A small, written data request. The initiative restated as workflow, metric, current value, target value, owner.
- Week 2
Baseline signed, thresholds drafted
The baseline measured with honest attribution — system-of-record pull first, time-and-motion sample second, self-report only to triangulate; every figure tagged [counted] or [estimated]. The sponsor signs the baseline as fair before any threshold is negotiated. The acceptance test is drafted, thresholds set against the signed number.
- Week 3
The test the client can run alone
The people who will run the test execute a dry run on a small sample with us observing; the exit condition is that they can run it without us. The scoped-pilot definition is finalized. A close-out read-out with the sponsor delivers the three artifacts, the assessment against the published protocol, and — where the honest answer is "don't run this pilot" — that answer, in writing.
what you get
Four things you keep, each one checkable by your own people. All are issued as an assessment against the published protocol — not a verdict, certification, or professional sign-off.
A written problem statement
The initiative rewritten as workflow, metric, current value, target value, and owner — the version that survives the rewrite, or the finding that it doesn't.
A sponsor-signed baseline
The target workflow's current-state metric: definition, measurement source, sample window, value, each figure tagged [counted] or [estimated], known confounders noted — signed by the sponsor as fair.
A vendor acceptance test on your data
Thresholds, measurement procedure, sampling rule, pass/fail definition, and dispute path — executable by your team with no Tektari involvement, usable verbatim as a contract exhibit with any vendor.
A Go/No-Go decision record
The three artifacts, the assessment against the published protocol, and a scoped-pilot definition with kill criteria — the decision, and where the answer is "don't run this," said in writing.
the price, and the fee at risk
$12,500 fixed. No hourly billing, no scope creep, no change orders inside the three weeks. The price lives on the site, not in a proposal.
- $10,000 Invoiced at kickoff, before Week 1.
- $2,500 Invoiced only when the scoped pilot passes its pre-agreed written acceptance test — never invoiced if the pilot never runs, and never invoiced if it fails.
The trigger is binary and client-verifiable: your team runs the acceptance test on your own data, and the pass/fail rule is applied mechanically to the numbers. There is no re-measurement bureaucracy, no partial credit, and no Tektari discretion in the trigger — that is what makes the tranche a real exposure instead of a marketing clause. Per the charter, this fee never credits toward any other Tektari offer.
what this is not
- — Not an implementation. No client-data ingestion, no tool procurement, no vendor introductions, no build. If the assessment says a pilot is worth running, we are not the ones who get paid to run it.
- — Not a readiness assessment, maturity model, or roadmap. Those are opinions; this ships a testable artifact your team can run without us.
- — Not a vendor recommendation. The acceptance test is vendor-neutral by construction — usable against any vendor's pilot, or an internal build. It does not warrant the vendor, the product, or its future performance.
- — Not professional advice. Deliverables are an assessment against the published protocol, for decision-support. They are not legal, financial, accounting, or engineering advice, and they don't replace the sign-offs your own experts owe you.
If the firm that grades your AI decision also sells the fix, the grade isn't independent.
The audit runs under Tektari's published Verification Charter: no implementation sold, no vendor commissions or referral fees, prices published, and verification fees never credit toward any other Tektari offer. Acceptance tests are written so your own team can run and verify them without us, on your own data — and part of our fee rides on that test. Teaching cases are clean-room synthetic; a live engagement uses only your own records.
See where you rank first.
The audit is the paid next step. Start with the free Industrial AI Scoreboard — twenty scored questions, no sales call — and see which band you're in before you spend a dollar.
Take the ScoreboardThat doesn't look like an email — try name@company.com