Skip to main content
04

Tool 04 · Production Readiness Ledger

Can you actually ship it?

Fourteen questions, each about an artefact that either exists or does not. No score out of 100, no maturity curve, no peer cohort. The output is GO, NOT YET or NO — with the exact missing artefacts named. If it says NOT YET, it tells you not to hire anyone, including us, until you have them.

Ninety seconds, nothing to estimate

Questions
14 artefact gates
Blocking
4 of them
Output
GO · NOT YET · NO
Score
None. Deliberately.

Every gate is an artefact, not an opinion, so there is nothing here you can flatter yourself into. Runs entirely in your browser: no email, no account, nothing stored.

01Why these gates

Gartner names three causes. These are the three column headings.

Over 40% of agentic AI projects are forecast to be cancelled by the end of 2027 — on escalating cost, unclear business value and inadequate risk controls. Every gate below sits under one of those three causes, because the cheapest way to avoid being in that 40% is to check whether the artefacts exist before anyone is paid to build anything.

PublishedGartner press release, June 2025(poll of 3,412)

Escalating cost

Nobody owns the number, the released time or the system a year from now.

3 gates

Unclear business value

Nothing to measure against, so nothing can be proven.

4 gates

Inadequate risk controls

No access, no authority, no way back.

7 gates

02The ledger

Answer for one workflow, not for the organisation

Pick the workflow you would actually put an agent on. Answer honestly — a don't know is counted as a no, because an artefact nobody can point to is not one you have. The verdict updates as you go and the link carries every answer.

Step zero — where the work sits

Inputs
V · Which vertical is this workflow in?

This swaps in the regulated pair of gates for healthcare and semiconductor. It changes nothing else — there is no industry weighting and no score to weight.

R · Is this workflow inside a regulated process?

When this is on, the retention and logging gate (G11) is replaced by a stricter pair: a reproducible reasoning trace, and the validation and change-control procedure the agent will be brought under.

Escalating cost

0 / 3 answered

Nobody owns the number, the released time or the system a year from now.

  1. G07 · A number the CFO already believes for what this workflow costs today.

    A cost figure that has survived finance once already. A number invented for the business case will be challenged exactly when you need it.

  2. G12 · A decision on what happens to the time the agent frees up, made by someone who owns that budget.

    A written answer to 'and then what?' from the person whose budget the released hours sit in. Without it, the saving is a slide.

  3. G13 · Somebody who will still own this system in twelve months.

    A named owner with a role that survives the project. An unowned agent is a liability with a monthly bill.

Unclear business value

0 / 4 answered

Nothing to measure against, so nothing can be proven.

  1. G01 · A written definition of what counts as a wrong answer for this workflow, agreed by the people who own the process.blocking gate

    A page naming the failure modes and, for each, what the correct outcome would have been — signed off by the process owner, not written by the project team alone.

  2. G05 · A measured baseline error rate for the humans doing this work today.

    A number, from a sample somebody actually checked, for how often the current process gets it wrong. Not an estimate from memory.

  3. G06 · At least 50 real historical items, with their correct outcomes, that could be used as an evaluation set.

    Fifty real cases with the right answer attached — the minimum from which an evaluation harness can be built.

  4. G14 · A test environment where a failure costs nothing.

    Somewhere the agent can be wrong repeatedly without touching a customer, a payment or a patient.

Inadequate risk controls

0 / 7 answered

No access, no authority, no way back.

  1. G02 · Read access to the system of record in a non-production environment, available today.blocking gate

    Working credentials against a non-production instance with representative data — not a promise that access can be requested.

  2. G03 · A named individual with authority to let the agent act without review, and the authority to stop it.blocking gate

    One person, named, who can both switch it on and switch it off. A committee is not an answer to this question.

  3. G04 · A rollback that does not require a vendor to execute.blocking gate

    A documented, rehearsed way for your own team to reverse what the agent did, without opening a support ticket.

  4. G08 · A written escalation path for the case the agent gets wrong — who is told, how fast, what they can do.

    A named recipient, a stated time, and a stated remedy. One page.

  5. G09 · Write access to the system of record, or a documented decision that the agent will only recommend.

    Either the write path exists, or somebody has decided in writing that it will not — both are fine, ambiguity is not.

  6. G10 · A named owner for the data the agent will read, who can grant access without a committee.

    One person who owns the data and can say yes.

  7. G11 · An agreed retention and logging policy that covers model inputs and outputs.

    A written policy stating what is logged, where it lives and how long it is kept — covering prompts and completions, not only application logs.

0 of 14 answered

VerdictProvisional — 0 of 14 answered

NO — not buildable yet

This cannot be built by anyone, including us, until G01, G02, G03 and G04 exist. They are free. They are also the 4 things that decide whether the build survives contact with production.

Blocking artefacts missing

4 named
  1. G01

    A written definition of what counts as a wrong answer for this workflow, agreed by the people who own the process.

    typically a day, once the right two people are in a room

  2. G02

    Read access to the system of record in a non-production environment, available today.

    typically a fortnight, and it is almost always the long pole

  3. G03

    A named individual with authority to let the agent act without review, and the authority to stop it.

    typically a day, and the difficulty is political rather than technical

  4. G04

    A rollback that does not require a vendor to execute.

    typically a few days to write down, longer to rehearse

Unanswered gates count as missing. An artefact you have not confirmed is not one you can rely on — but the verdict updates the moment you answer.

We would not take this build yet either. Thirty minutes on closing the blockers — no pitch. The blockers below are internal work, and they are free. That half hour is worth more to you right now than a proposal.

Book thirty minutes

We reply within one business day.

Rule 1 — any blocking gate missing returns NO. Rule 2 — 3 or more non-blocking gates missing returns NOT YET. Rule 3 — everything else returns GO. There is no score, no maturity curve and no peer cohort. See the full rubric →

The one gateG05

You cannot state the error rate of the humans doing this work today. That means you will have no way to prove the agent is better, and no way to defend it the first time it is wrong.

A number, from a sample somebody actually checked, for how often the current process gets it wrong. Not an estimate from memory.

typically a fortnight of sampling, and it is the single most valuable fortnight in the project

The ledger

Every gate, its cause group, and where you stand

In place
0
Missing
14
Don't know
0
Gates
14
The readiness ledger: every gate, grouped under Gartner's three named causes of agentic project cancellation, with the artefact definition and typical internal effort for each gate that is missing.
GateArtefactStatus
Escalating cost
G07

A number the CFO already believes for what this workflow costs today.

A cost figure that has survived finance once already. A number invented for the business case will be challenged exactly when you need it.

typically a week

Not answered
G12

A decision on what happens to the time the agent frees up, made by someone who owns that budget.

A written answer to 'and then what?' from the person whose budget the released hours sit in. Without it, the saving is a slide.

typically a day, and it is usually the conversation nobody wants to have

Not answered
G13

Somebody who will still own this system in twelve months.

A named owner with a role that survives the project. An unowned agent is a liability with a monthly bill.

typically a day

Not answered
Unclear business value
G01Blocking

A written definition of what counts as a wrong answer for this workflow, agreed by the people who own the process.

A page naming the failure modes and, for each, what the correct outcome would have been — signed off by the process owner, not written by the project team alone.

typically a day, once the right two people are in a room

Not answered
G05

A measured baseline error rate for the humans doing this work today.

A number, from a sample somebody actually checked, for how often the current process gets it wrong. Not an estimate from memory.

typically a fortnight of sampling, and it is the single most valuable fortnight in the project

Not answered
G06

At least 50 real historical items, with their correct outcomes, that could be used as an evaluation set.

Fifty real cases with the right answer attached — the minimum from which an evaluation harness can be built.

typically a week, assuming the outcomes were recorded at the time

Not answered
G14

A test environment where a failure costs nothing.

Somewhere the agent can be wrong repeatedly without touching a customer, a payment or a patient.

typically a fortnight

Not answered
Inadequate risk controls
G02Blocking

Read access to the system of record in a non-production environment, available today.

Working credentials against a non-production instance with representative data — not a promise that access can be requested.

typically a fortnight, and it is almost always the long pole

Not answered
G03Blocking

A named individual with authority to let the agent act without review, and the authority to stop it.

One person, named, who can both switch it on and switch it off. A committee is not an answer to this question.

typically a day, and the difficulty is political rather than technical

Not answered
G04Blocking

A rollback that does not require a vendor to execute.

A documented, rehearsed way for your own team to reverse what the agent did, without opening a support ticket.

typically a few days to write down, longer to rehearse

Not answered
G08

A written escalation path for the case the agent gets wrong — who is told, how fast, what they can do.

A named recipient, a stated time, and a stated remedy. One page.

typically a day

Not answered
G09

Write access to the system of record, or a documented decision that the agent will only recommend.

Either the write path exists, or somebody has decided in writing that it will not — both are fine, ambiguity is not.

typically a fortnight if write access is needed

Not answered
G10

A named owner for the data the agent will read, who can grant access without a committee.

One person who owns the data and can say yes.

typically a day to identify, longer if the answer is nobody

Not answered
G11

An agreed retention and logging policy that covers model inputs and outputs.

A written policy stating what is logged, where it lives and how long it is kept — covering prompts and completions, not only application logs.

typically a week with legal

Not answered

The link carries every answer. No email, no gate, nothing stored on our side.

03How this is judged

Every rule, printed before you answer

No competitor in this landscape shows its weights. We have none to hide: there is no arithmetic here, no coefficient and no score — only a list of gates and three rules applied in order.

The verdict rules

  1. 01Any blocking gate (G01–G04) answered No or Don't know.no
  2. 02Three or more non-blocking gates answered No or Don't know.Not yet
  3. 03Everything else.go

A don't know counts as a no for the verdict, and is reported separately and prominently. The threshold in rule 2 is 3 — it is not tuned, it is published, and you are welcome to argue with it.

The 4 blocking gates

These four are not weighted more heavily. They are absolute: any one of them missing returns NO on its own, because without them nobody can build this — including us.

  • G01A written definition of what counts as a wrong answer for this workflow, agreed by the people who own the process.
  • G02Read access to the system of record in a non-production environment, available today.
  • G03A named individual with authority to let the agent act without review, and the authority to stop it.
  • G04A rollback that does not require a vendor to execute.

In a regulated process, G11 is replaced by a stricter pair — G11a and G11b. Nothing else changes.

The gate list and these rules are one pure module that both this page and the tool import, so the disclosure cannot drift from the code that produces the answer. Read the full methodology →

04Limits

What this does not measure

Transparency is the only thing we can honestly substitute for scale. So here is what this instrument is not.

  • This is our rubric, not an industry benchmark. Every gate and every verdict rule is published above and on the methodology page.

  • It scores production-readiness for one workflow, not organisational AI maturity. Those are different questions and conflating them is how readiness quizzes became worthless.

  • We deliberately do not place you against a peer cohort, because we do not have one. Cisco can, with 7,985 double-blind respondents across 30 markets. Borrowing that language without the sample would be the fastest way to be dismissed.

Not sure which of these you need? Start with the necessity test → Or go back to the toolkit →