Process

How Statepath keeps the route finite, legible, and low theatre.

The process is designed to make the current state readable, keep access proportionate, and return one decision-ready route without blurring into a much larger engagement.

Access Read-only preferred where possible.
Scope Agreed surfaces only.
Closeout Temporary access removed at closeout.

Working sequence

The route should feel deliberate, not chaotic.

Each step exists to keep scope tight and the result decision-grade.

01

Short fit call

Check there is a real project, a real confusion point, and a bounded reason to investigate before anything heavier starts.

02

Proposal and scope lock

Agree exactly what will be reviewed and what stays outside the audit so the work does not quietly sprawl.

03

Access and onboarding

Use the minimum practical access route, with read-only access preferred where possible.

04

Audit and review

Inspect the agreed surfaces, separate what is direct, derived, or unresolved, and assemble the decision pack without blurring those truth states together.

05

Return one next route

Deliver one sensible next move. If a bounded corrective lane is found, it can be quoted separately. If the truth points to something broader, the broader route is named explicitly instead of blurred.

What gets checked

The audit checks the agreed surfaces, their contradictions, and the decision risk around them.

Process is not just sequence. It also defines what Statepath is actually looking for inside the named repo, docs, dashboards, automations, and handover material.

Surface group one

Repo, code, and implementation surfaces

Review the parts of the live build, scripts, configuration, or automation logic that are actually in scope for the decision.

Surface group two

Docs, handover notes, and operating story

Check whether the current explanation still matches the current implementation or whether the narrative has drifted from the truth on disk.

Surface group three

Dashboards, flows, and live-route behaviour

Inspect the named dashboards, integrations, exports, or flow surfaces that matter to the actual decision instead of assuming the route lives only in code.

Surface group four

Contradictions, blockers, and decision risk

Look for the specific gaps, conflicts, or fragile assumptions that would make the next wrong move expensive if they remain unresolved.

What this prevents

The method is designed to stop three predictable failures before they compound.

Premium process should feel safer because the route blocks avoidable sprawl, over-access, and confidence-heavy guesswork before they become expensive.

Failure one

Scope drift before truth exists.

Without a fixed sequence, the work quietly expands from one bounded audit into broad rescue labour before the baseline has even been made legible.

Failure two

Too much access before real need is proven.

The process keeps access proportionate so the review can stay serious and bounded instead of defaulting to “give us everything” by habit.

Failure three

A recommendation that outruns what the surfaces actually prove.

The sequence forces confirmed truth, caveats, and one next route into the same pack so the answer stays decision-grade rather than confidence-heavy.

Truth labels

The pack keeps direct, derived, and unresolved material separate on purpose.

This is the discipline that stops a tidy summary from over-claiming what the agreed surfaces can really prove.

Label one

Direct

Use direct when the agreed surfaces themselves clearly support the claim without extra interpretation doing the real work.

Label two

Derived

Use derived when the route needs reasoned interpretation across multiple signals, even if that interpretation is sensible and well-supported.

Label three

Unresolved

Use unresolved when the current surfaces cannot prove the point cleanly enough yet and caution still belongs in the recommendation.

Trust and privacy

Operate with the minimum access needed, and keep the route inspectable.

Minimum access, agreed surfaces, and a clean closeout matter more than looking busy.

Access posture
  • read-only access is preferred where possible
  • only the agreed in-scope materials should be used
  • temporary access is preferred to shared personal credentials
  • access should be removed at closeout
AI-assisted work posture

Client secrets, credentials, private code, sensitive personal data, or confidential commercial material should not be placed into third-party AI tools unless explicitly agreed in writing for that engagement.

Best first note

The strongest first contact is short and concrete.

The goal is enough context to judge fit, not a novel.

After the audit

If one clearly bounded corrective route exists, it can move forward cleanly.

That follow-through lane is quoted separately at £1,500 fixed. If the truth points to something broader, the broader route is re-scoped explicitly rather than being smuggled into the original engagement.

Output delivery boundary

The first delivery ends with a pack and one bounded route, not a hidden implementation commitment.

This is where the process turns into a usable handover: clear enough to act, but still honest about what belongs inside the first step and what remains separate.

Included in delivery

One decision-ready audit pack

  • first-read summary
  • current-state truth map
  • risk and caveat notes
  • one bounded next-route recommendation
Separate by design

Anything broader than the first step

  • open-ended implementation labour
  • permanent support expectations
  • broad rescue work hidden inside the audit
  • follow-through unless one bounded corrective lane is clearly justified
If one clearly bounded corrective lane exists, optional follow-through can be quoted separately. If not, the broader route must be re-scoped explicitly rather than implied by the audit pack.

Linked pages

Use the page that answers the next question instead of forcing everything into one slab.

That is the whole point of the new structure.