- The Checklist We Run Internally
The 12-Point AI Readiness Checklist
This is the checklist we work through in the first days of every assessment. It is published in full because a business that runs it honestly and finds nothing is a business we should not be selling to.
How to Use It
Four Categories, Twelve Questions, One Honest Score
Work through the twelve items below and mark each one yes, partly, or no. Answer as your business actually is today, not as it will be once the project everyone keeps mentioning finally finishes.
Nine or more clear yeses means you are ready to scope real work. Between five and eight means you are ready for a narrow pilot in the area where your answers were strongest. Below five means the foundation work will pay back faster than anything AI can do for you right now, and it is cheaper.
We work through this checklist in person with Dallas-area owners, driving down from Plano rather than emailing a form to fill out. Most of the useful answers surface in the room, usually from the office manager who knows exactly which report nobody trusts.
Category One, Data
1
Can you name where your most important data lives?
Not roughly. Specifically, which system, which fields, who administers it.
2
Is that data consistent enough to trust?
Duplicate records, three spellings of one client name, and blank required fields all produce confidently wrong output.
3
Can a tool reach it without a person exporting a spreadsheet?
If every workflow starts with a manual export, automation ends where that person’s attention does.
Category Two, Process
4
Is the target process written down?
If it exists only in one experienced person’s head, that has to be fixed first, and it is worth doing regardless.
5
Would two experienced staff do it the same way?
Variation between capable people means the rules were never actually agreed, only inherited.
6
Does it run often enough to matter?
Many times a week, not many times a year. Volume is what turns a small time saving into a real number.
Category Three, Security & Compliance
7
Do you know who can access what?
The most common blocker we find. AI tools that surface documents expose permission drift immediately, usually in front of staff.
8
Do you have a written AI use policy?
9
Do you know which data is legally restricted?
Patient records, client files, donor data. If the answer is “probably some of it,” start with the regulated pathway (linked text).
Category Four, People
10
Is there a named owner for the outcome?
One person, named, who would be asked what happened, and who can change the process rather than just encourage people to follow it.
11
Does the team have capacity to absorb a change?
A team already at capacity quietly reverts within a month, and the license renews unused.
12
Can you state what success looks like as a number?
Hours saved, response time reduced, errors caught. If you cannot name it before starting, you cannot judge the result after.
Scoring it honestly is the hard part
Most businesses mark item 7 as a yes and then discover during the assessment that sharing permissions drifted years ago. If you are unsure on any item, treat it as a “partly.” The whole value of this exercise is finding out before you spend, not after.
We scored low. Is that unusual?
Not remotely. A significant share of businesses that ask us about AI score below the pilot threshold on first pass, and almost all of them are running fine as businesses. A low score is a statement about foundation work, not about how well you are run.
Which items matter most?
Seven and ten. Unknown permissions create the fastest and most embarrassing failures, and no named owner is the most reliable predictor of a project that quietly stops. Items one through six are fixable in a defined amount of work; those two are cultural.
If we can fix these ourselves, why hire you?
If you can, you should. Plenty of businesses run this checklist, find two gaps, close them internally, and come back a quarter later ready. That is a better outcome than paying us to tell you the same thing more formally.
How is this different from your full assessment?
This is the screening layer. The assessment verifies each answer against the actual systems rather than accepting the self-report, maps the processes properly, and produces the ranked roadmap. The checklist tells you whether that is worth commissioning yet.
BK
Last reviewed August 2026 · Questions we have not answered here? Call 214-444-3583. Kehr Technologies is based in Plano, Texas, and works with businesses throughout the Dallas area.