Take our free AI readiness assessment: 15 questions across five dimensions, scored out of 30, telling you whether to fix, test or build before you spend on AI.
An AI readiness assessment tells you whether your business can get a return from AI right now, and what to fix first if it can’t. The quickest honest version is the scorecard below: five dimensions, three questions each, scored 0, 1 or 2, for a total out of 30. It takes about 20 minutes.
TL;DR: Score the 15 questions from 0 to 2 and add them up. Under 11, fix your processes and data before buying anything. From 11 to 20, test one well-chosen workflow. Above 20, you are ready to build.
What an AI readiness assessment should tell you
Most published readiness assessments are maturity models. They rate you from “Level 1: Ad hoc” to “Level 5: Optimised” without saying what to do on Monday. A business with 15 to 200 people needs a narrower answer: is there one workflow where AI would save real hours or money, and are the conditions in place to make it work? A useful AI readiness checklist answers that, and is willing to say “not yet”.
The SME AI Pulse, a monthly survey commissioned by the National AI Centre and run by Fifth Quadrant, found 43% of Australian SMEs reported some level of AI adoption from December 2025 to February 2026, a slight dip from 45% the quarter before. Trust was the biggest barrier: around 65% of non-adopters cited distrust in AI decision-making or a strong preference to keep human control. Using a tool is not the same as getting a return from it.
For the governance questions, this scorecard borrows from the government’s Guidance for AI Adoption, released by the National AI Centre in October 2025 and known as AI6. Its six practices are: decide who is accountable, understand impacts and plan accordingly, measure and manage risks, share essential information, test and monitor, and maintain human control. The guidance is voluntary and about responsible use, so I have added the commercial questions it leaves out. A business can follow every practice and still waste its budget.
The scorecard: 15 questions, five dimensions
If you sit between two descriptions, take the lower one. Have the person who does the work score it as well as the person who manages it, and compare.
Dimension 1: Process clarity
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 1. Can you name the specific workflows that consume the most staff hours each week? | We can’t say which workflows take the most time | We can name them, but the hours are estimates | Hours per workflow are tracked, or can be pulled from timesheets or system logs |
| 2. Is each of those workflows documented, step by step? | It lives in one person’s head | Partly written down, or out of date | Written down, current, and a new starter could follow it |
| 3. How consistently is the work done? | Every person does it their own way | Broadly the same, with frequent unplanned exceptions | Standardised, with written rules for the exceptions |
Dimension 2: Data and systems
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 4. Where does the data for your key workflows live? | Mostly in spreadsheets, inboxes and paper | Split between proper systems and spreadsheets | Mostly in one or two systems of record |
| 5. How clean is that data? | Duplicates and gaps mean reports can’t be trusted | Usable, but someone cleans it by hand before each report | Reports run straight from the system, and a named person owns data quality |
| 6. Can your systems share data with other tools? | No exports, no integrations | Manual exports or CSV files only | Documented APIs, or built-in connectors to tools such as Microsoft 365, Xero or your CRM |
Dimension 3: People and skills
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 7. How do staff use AI tools today? | Not at all, or we don’t know | Some staff use tools like ChatGPT or Copilot on their own initiative, with no approved tools | The business provides approved tools and most staff use them for real work |
| 8. Does someone own AI and automation internally? | Nobody | Someone does it as a side task | A named owner with time set aside |
| 9. How do staff feel about AI changing their work? | Fear or open resistance | Mixed: some keen, some quietly worried | Curious, and involved in choosing what gets automated |
Dimension 4: Governance and risk
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 10. Do you have a written AI use policy? | None | Informal rules or a verbal “be careful” | Written, shared with staff, reviewed at least yearly |
| 11. Do staff know what personal or client-confidential data they may put into AI tools? | No rule exists | A rough rule, not written down or enforced | Written rules on which data may go into which approved tools, checked against OAIC guidance |
| 12. Is a human accountable for checking AI output before it reaches a customer or a ledger? | No check | Checked sometimes, depending on who did the work | A defined review step and a named accountable person |
Dimension 5: Commercial case
| Question | 0 | 1 | 2 |
|---|---|---|---|
| 13. Can you put a dollar value on the problem you want AI to solve? | No, it just feels slow | A rough estimate | Costed in hours and dollars per month |
| 14. Do you have a budget and an approver for a first project? | None | Discussed but not allocated | Budget approved, or an approver has committed |
| 15. Have you defined what success looks like for a first project? | No | A general goal such as “save time” | A number and a date, for example “cut processing from 12 minutes to 3 by the end of the quarter” |
What your AI readiness score means
Add your 15 answers. The maximum is 30.
| Score | Band | What it means | What to do next |
|---|---|---|---|
| 0-10 | Not ready yet | The foundations are missing, and AI will amplify the mess | Fix process and data first; do not buy AI software |
| 11-20 | Ready to test | Enough foundation to prove value on one workflow | Pick one workflow and run a scoped audit or pilot |
| 21-30 | Ready to build | Conditions are in place for a real implementation | Scope the build, pick tools and set success metrics |
0-10: Not ready yet
Buy nothing yet. A chatbot on top of undocumented processes automates the confusion, and an AI agent reading data split between spreadsheets and someone’s inbox produces confident nonsense.
The fix mostly costs staff time. Document your two or three biggest workflows. Pick a system of record for customer and financial data and clear out the duplicates. Write a one-page AI use rule so staff stop guessing. Our post on the data problem no one talks about before implementing AI covers the data side, and legacy systems and AI adoption deals with core systems that get in the way. Rescore in three months.
11-20: Ready to test
The usual mistake here is picking the exciting workflow rather than the valuable one. A middle score means you can succeed on something narrow, not that the business is ready for wholesale change.
What you need is a short, scoped piece of work that proves one workflow. That is what our AI Quick-Win Audit is for. It is a fixed $2,950 fee: we map one of your workflows, identify the automation opportunities in it, and show you a live working demonstration. It starts with a free 30-minute fit call, and if you commission a build of $10,000 or more within 3 months, the fee is credited toward it.
If you would rather run the test yourself, follow our 90-day AI pilot guide, and read why AI pilots fail before you start.
21-30: Ready to build
Above 20, another assessment won’t tell you much. You need a scope, a delivery team and a success metric.
High scorers tend to trip on two things. The first is over-scoping: ship the workflow with the clearest dollar value before starting a second. The second is dropping the review step after launch. AI6 includes “test and monitor” and “maintain human control” because model behaviour changes and edge cases turn up in production. Document-heavy, rules-based work such as invoice and accounts payable automation makes a good first build, and our AI and automation services page explains how we scope and deliver it.
What an outside assessment adds
Self-scoring is free, but you are marking your own homework. An outside view adds three things.
The first is process mapping. The documented version of a workflow is often not how the work actually gets done, and sitting with the person doing it exposes the re-keying and workarounds where most of the savings hide.
The second is comparison. Question 13 asks you to cost one problem; an outside view ranks several candidates on hours saved, difficulty and risk, so the first project is the best return rather than the one with the loudest internal champion.
The third is a demonstration on your own data. It moves the conversation from “could AI do this?” to “here it is, and here is what it still gets wrong”. For longer-range planning, see our AI strategy and roadmapping service.
I have spent more than 20 years in technology, strategy and delivery, including leading digital transformation programs for NSW Government, UNSW Sydney, Telstra and Woolworths. My view from that work is that readiness comes down to process and ownership far more than to which AI tool you pick.
If you can’t tell which band you are in, the free fit call will sanity-check it, or you can contact us directly. For questions 10 to 12, our AI policy template for Australian businesses gets a written policy in place quickly.
Privacy and vendors
Questions 11 and 12 are easy to overscore. The Office of the Australian Information Commissioner recommends, as best practice, that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools. If staff paste client files into free chatbots, score question 11 as 0 whatever your policy says. When you choose a vendor, what to ask an AI vendor before you sign lists the questions.
Where to go from here
Run the scorecard this week with the person who does the work and the person who manages it. If you land in the middle band and want the fastest route from score to working demonstration, book the fit call for the AI Quick-Win Audit. An honest AI readiness assessment saves more money by stopping a bad purchase than by starting a good one.
Frequently Asked Questions
What is an AI readiness assessment?
An AI readiness assessment is a structured check of whether a business has the processes, data, people, governance and commercial case to get a return from AI. A useful one ends with a clear next step rather than a long maturity report.
How do I score my AI readiness?
Answer the 15 questions in the scorecard above, scoring each 0, 1 or 2, then add them up for a total out of 30. A score of 0-10 means fix your foundations first, 11-20 means test one workflow, and 21-30 means you are ready to build.
What should be on an AI readiness checklist?
Cover process clarity, data and systems, staff skills and ownership, governance and privacy, and a costed business case with a success measure. Check governance most carefully, because it is easy to assume a verbal rule counts as a policy.
Is my small business ready for AI if my score is low?
Not yet, and that is a useful result. A low score usually points to undocumented processes and scattered data, which take staff time to fix rather than new software. That groundwork gives any later AI project a far better chance of paying back.
How much does a professional AI readiness assessment cost?
It varies by provider and scope. Our AI Quick-Win Audit is a fixed $2,950, starts with a free 30-minute fit call, and the fee is credited toward any build of $10,000 or more commissioned within 3 months.