AI Readiness Assessment
Most agent projects fail on data access, governance and change capacity rather than on the model. This assessment asks twelve factual questions across those four areas and scores them out of 36. Every question asks what is true today, not how mature you feel, because a score built on self ratings measures optimism. It takes about ten minutes and nothing is transmitted while you answer.
Updated August 5, 2026Free, no signupRuns in your browser
Question 1 of 12
Data and systems
Twelve questions, about ten minutes. Your answers stay in this browser until you choose to send them.
Methodology and assumptions
What the model does, and what it refuses to guess
Twelve questions, four options each, scored zero to three. Thirty six points in total, spread evenly across four categories of three questions.
Why these four categories
Data and systems covers EHR API access, the reliability of the structured data an agent would read, and who would build and hold the integration. This is the category that most often turns a six week pilot into a six month one, because interface work sits in a shared queue you do not control. The EHR pages set out what the major platforms actually expose.
Governance covers whether a written AI policy exists, who signs off a clinical deployment, and how an AI vendor's risk would be assessed. Governance is what lets a pilot become production. Without it every deployment decision escalates and most of them stall. The controls themselves are shaped by the HIPAA Security Rule and, increasingly, by state AI statutes that differ by jurisdiction.
Operations covers workflow documentation, baseline measurement and change capacity. An undocumented workflow cannot be automated, only guessed at, and without a baseline you cannot prove the agent helped.
Strategy covers executive sponsorship, budget and use case clarity. A named sponsor and an approved budget line are the difference between a project and an experiment.
How the scoring works
Each option is a factual statement about your organization, ordered from least to most ready and scored zero to three by position. Every option is checkable. We have not asked you to rate your governance maturity out of five, because an instrument built on self ratings measures confidence rather than readiness, and the two are frequently inverted.
There is no weighting. The four categories carry nine points each, because in practice any one of them at zero will set your timeline regardless of the other three. That is also why the per category breakdown matters more than the total, and why two organizations on the same score can be in very different positions.
The bands are: zero to twelve Exploratory, thirteen to twenty two Emerging, twenty three to thirty Ready for a bounded pilot, and thirty one to thirty six Ready to scale.
Why there are no numeric inputs
This is an instrument rather than a calculator, so it carries no editable defaults. The twelve questions, the four bands and the per band next steps are the interactive body of the tool. If you want a model you can put your own numbers into, the AI scribe ROI calculator and the prior authorization cost calculator both take your figures and print every formula.
What this is not
It is a structured self assessment, not an audit. It is only as accurate as the answers and it checks nothing against your systems. A real readiness audit verifies the API access, reads the data and interviews the people doing the workflow, which routinely produces a lower score than the one people expect. Treat this as a way to find the weakest link and to have a more specific conversation, not as a certificate.
Nothing is sent anywhere while you answer. The scoring runs in your browser. Your band and score are shown to you for free the moment you finish. Only the per category breakdown and the tailored next steps ask for an email, and that is so we can send you a copy.
Questions we get asked
How long does the AI readiness assessment take?
About ten minutes if you know the answers, and longer if you do not, which is itself informative. Several questions have an honest answer of we do not know, and those are usually the most useful ones to take away from it.
Do I have to give an email address to see my score?
No. Your band, your score out of 36 and the summary for that band appear as soon as you finish, with no form. The per category breakdown and the tailored next steps sit behind a short form, because that is the part we email you a copy of. We say which half is which before you answer anything.
Who should fill this in?
Ideally not one person. The data and systems questions are usually answered accurately only by IT, and the governance questions only by whoever owns risk or compliance. Filling it in as a group takes half an hour and produces a far more honest score than one person guessing on behalf of three departments.
What is a good score?
Look at the lowest category before the total. Twenty four points made up of six in each category is a very different position from twenty four points with nine, nine, six and zero, and the second one will be blocked by whatever scored zero regardless of the rest. In our experience the category that most often sits lowest is governance, and it is also the cheapest one to fix.
Is this the same as your AI Readiness Audit?
No. This is a self assessment you score yourself in ten minutes. The audit is an engagement in which we verify what your EHR actually exposes, look at the data and interview the people doing the workflow. The self assessment is a good way to decide whether the audit is worth commissioning, and to know which category to point it at first.
What should we do with a low score?
Fix the lowest category first, because it is the binding constraint. A low governance score is usually addressable in weeks through a written policy, a risk tiering and a named owner. A low data and systems score is slower, since interface work depends on your vendor's queue, and knowing that early changes which use case should go first.
What happens to my answers?
The scoring runs entirely in your browser and nothing is transmitted while you answer. If you choose to unlock the full breakdown, your name, work email, organization and the score summary are sent to us so we can email you the copy. We do not sell or list build with it, and there is a business associate agreement in place before any conversation that touches patient data.
Make it a formal evaluation
Everything we publish is free to read and free to argue with. When the decision has to be signed, dated and defended to a board, we run the evaluation against your own estate. We take no vendor commissions.
- A 30 minute evaluation call with an analyst, no pitch deck.
- A read on the vendors and the rules in play, and the use cases we would not touch yet.
- A written proposal with scope, sequence and a fixed fee.
- No obligation
- Direct with an analyst, not a sales rep
- BAA available before any PHI discussion