Skip to content
Orbav
Orbav
Book a build reviewSee the Sprint

The AI Readiness Audit

Twelve questions. One email field. A scored report on where AI and automation would actually pay in your business — and, just as usefully, where they wouldn't.

01

What you get

A scored report across four dimensions:

Data readiness
Whether your data is in a state anything could be built on.
Process automation potential
Which processes are worth automating and which need fixing first.
AI use-case fit
Where a model genuinely helps, and where it’s an expensive way to do something simple.
Risk and compliance posture
What would need handling before anything goes near production.

Plus your top three opportunities, ranked by likely payback.

02

What it looks like

Below is the report’s real output — the same scoring code and the same document you’ll get, run on a fixed set of example answers. Not a mockup, and no one’s business but the maths’.

Worked example · synthetic answers · captured 2026-08-13 · same document the report email links to

Read the full worked example

03

Get the audit

Pick the closest answer on each — there’s no perfect one, and “not sure” usually means the second-lowest option is true.

01 · Data readiness

Whether your data is in a state anything could be built on.

1.Where does the data behind the process you’d automate actually live?
2.When two systems disagree about the same customer or order, what happens today?
3.Could you export the records a build would need — customers, deals, tickets — without chasing anyone?

02 · Process automation potential

Which processes are worth automating and which need fixing first.

4.Is there a process your team repeats the same way, at least weekly — moving data between tools, chasing approvals, copying updates?
5.Do those steps exist in writing anywhere outside someone’s head?
6.When that process goes wrong today, how do you find out?

03 · AI use-case fit

Where a model genuinely helps, and where it’s an expensive way to do something simple.

7.Does the work involve reading messy, free-text input — emails, documents, requests — before anyone can act on it?
8.For the task you have in mind, is “about right” ever acceptable — or must the output be exactly correct?
9.What has already been tried for this task?

04 · Risk and compliance posture

What would need handling before anything goes near production.

10.Are there rules about where your customers’ data can go — contracts, GDPR, sector regulation, client terms?
11.Before something automated touches a customer record or money, who signs off?
12.If an automated step made a mistake in production tomorrow, could you see it, stop it and undo it?

Answer all twelve questions — about six minutes.

required

The report lands here. No phone number, no company-size dropdown, no ‘how did you hear about us’ — the five follow-up emails are part of the deal, and they stop the moment you reply STOP.

04

What happens next

The report arrives immediately. Then five emails over two weeks, each one a specific thing we've learned about making AI systems that don't fall over. You can unsubscribe from any of them, and we don't call you because you downloaded something.

Your email, your twelve answers and the scores they produce are stored in our own database table under the same rules as contact messages — kept at most 24 months, deletable on request. The privacy policy has the full text.

05

FAQ

Is it really free?
Yes, and there’s no call attached.
How long does it take?
About six minutes.
Will you sell my email?
No. Sub-processors are listed on the Trust page.

06

Or start somewhere concrete

If you already know which process is costing you money, the Sprint is a faster answer than an assessment.

See the Sprint