AI Readiness Audit
Find out what is worth building before you pay to build it
Two weeks. $2,400 fixed. We map how the work is really done, test whether your systems can actually be integrated — against a live API, not a brochure — and hand you a ranked, costed build plan you own. Including, where it is the honest answer, the recommendation not to build yet.
- Fixed. Not a starting point
- $2,400Fixed. Not a starting point
- Start to written plan
- 2 weeksStart to written plan
- Named, and yours to keep
- 5 artefactsNamed, and yours to keep
- Of your team’s time, total
- ~6 hoursOf your team’s time, total
Request the audit
Five questions, then a real answer
We ask for the process and the systems because that is what decides whether two weeks is enough. It means the first reply can tell you something useful rather than ask you for the same details again.
What happens next
- Day 0
You send this form
We read it and reply within one business day, either with a slot or with the reason an audit is not what you need yet. Both answers are free.
- Day 1–2
A 45-minute scoping call
We agree which processes are in scope, who we need 30 minutes with, and which systems we will need read-only credentials or a sandbox for.
- Week 1
We watch the work and read the systems
Short sessions with the people who do the process, then the integration test — we authenticate and attempt the reads and writes a build would actually depend on.
- Week 2
Five artefacts and a walkthrough
The process map, the feasibility report, the ranked shortlist, the build plan and the risks — delivered, then walked through live for an hour with your team.
Would rather talk first?
Thirty minutes, no preparation and no deck. We will tell you on the call whether the audit is the right starting point for you.
Book a discovery callWhat you receive
Five artefacts, named before we start
Five artefacts, named up front. They are written for two audiences at once — an executive who needs the decision, and an engineer who needs to check the reasoning — and they are yours whether or not you build with us.
- 01
Process map of how the work is really done
Every step of the processes in scope, the system each step touches, who does it, how long it takes and how often it happens. Drawn from watching the work, not from your process documentation — the two are rarely the same, and the difference is usually where the automation case lives.
Step-level map with timings and volumes
- 02
Integration feasibility report, tested not assumed
We authenticate against your API, sandbox or database and try the reads and writes an agent would actually need. Rate limits, missing endpoints, undocumented behaviour and permission models are found now, for $2,400, rather than in week five of a build.
Per-system verdict with the evidence from the test calls
- 03
Ranked shortlist with hours and cost against each process
Each candidate process scored on volume, rule clarity, integration difficulty and the cost of being wrong — with the current hours it consumes and an estimated build cost beside it. Ranked, not listed, so there is a defensible first project rather than a menu.
Scored table, top three argued in prose
- 04
Written build plan
For the recommended first project: the architecture, which systems are read and which are written to, what scope each credential carries, which decisions stop for a person, the phases, and a cost and timeline range. Specific enough to be quoted by somebody else.
Build plan you can take to any vendor
- 05
The risks and the case against
What could make this fail: data quality, a system that cannot be integrated, a process that should be redesigned rather than automated, or a team that will work around it. Including, where it applies, the recommendation to do nothing yet.
Named risks with a mitigation or a stop
Before you send it
Who this is and is not for
An audit that cannot get access produces a document nobody can act on. It is worth thirty seconds to check that the right-hand column does not describe you.
A good fit when
- You have a process that runs often enough to be worth automating, and people who can describe it.
- The systems involved have an API, a database or a sandbox we can be given read access to.
- Somebody can sign off on roughly six hours of your team’s time across two weeks.
Not the right starting point when
- You want a maturity score against a framework. We do not produce one — the audit tests feasibility instead.
- You already know the process and the systems and just want it built. Go straight to the Agent Pilot.
- Nobody internally can spare time for the sessions. The audit is only as good as the access it gets.
The full package page has the method, the timeline and the published not-included list.
Questions
Asked before every audit
- How much does the AI Readiness Audit cost?
- $2,400, fixed. Not a starting point and not an estimate — the figure does not change with what we find, and there is no obligation to build anything afterwards.
- How long does the audit take?
- Two weeks from the scoping call to the written build plan, including the hour-long walkthrough of the findings at the end.
- How much of my team’s time does it take?
- About six hours in total across two weeks: a 45-minute scoping call, roughly 30 minutes each with three or four people who do the work, access to a system or sandbox, and a 60-minute walkthrough at the end.
- What do I actually receive?
- Five artefacts you own outright: a step-level process map with timings and volumes, a tested integration feasibility report, a ranked and costed shortlist of candidate processes, a written build plan for the recommended first project, and a named list of risks. They are written to be quoted by any vendor, not only by us.
- What if the audit says we should not build anything?
- Then that is what the document says, in writing, with the reasoning. It is a real outcome rather than a polite one — and you still keep the process map and the feasibility report, which is usually the more valuable half.
- Do I have to build with Devs Core afterwards?
- No. The build plan is specific enough to be quoted by someone else, which is deliberate. If you take it elsewhere the audit has still done its job.