Mac Sweeny
Quick answer
An AI readiness assessment checks whether a specific workflow has the inputs, people and review process needed for a useful trial. Its output should be a decision and a next step, supported by evidence. A high score alone does not prove a return.
Three checks before you start
Use these questions to decide whether a workflow is ready for a small trial.
- 01
Is the task clear?
Can you explain the inputs, output and what a good result looks like?
- 02
Are the inputs usable?
Are the records reliable and appropriate to share with the chosen tool?
- 03
Who will review it?
Name the owner and decide how to spot and handle a bad result.
An unclear answer is a reason to fix that part of the workflow first.
Put it to work · Added 7 September 2026
Is one workflow ready for a small trial?
Assess a specific workflow, not your whole business with one score. You need an outcome, representative inputs, an owner, a reviewer, a baseline and a way to stop the trial when it fails.
I’d use these checks to find the next piece of preparation. Ticking every box is a starting point for a supervised trial, not a certification or a prediction of return.
| Area | Evidence you should have | If it is missing |
|---|---|---|
| Outcome and baseline | A named task, current handling time and a measure of acceptable quality | Observe the work before choosing a tool |
| Inputs and access | Representative examples you are permitted to use in the proposed tool | Resolve access and permitted use before uploading anything |
| People and review | An owner, a reviewer and examples of exceptions | Name the people and explain how they will check outputs |
| Trial and stop rule | A small test, a comparison method and a manual fallback | Agree when to pause and how work continues |
Check your preparation
Start with any missing evidence. This checklist has no pass mark.
These controls work in this page. They do not submit your answers or save them for later.
Your next step. Choose the first unchecked item and resolve it with the process owner. When the basics are in place, trial with one or two users and compare time, corrections and usability before expanding.
Download the worksheet (CSV)What an AI readiness assessment actually covers
I’d look at the outcome, the work as it happens today, the information available and the people who will own it. Then I’d compare options, including changing the process or using a feature you already have.
The worksheet above is a planning aid. It is not a legal, security or compliance assessment. Resolve any specialist question with the appropriate person before relying on the proposed workflow.
Run the DIY version: ten questions
- Which repeated task are we trying to improve?
- How often does it happen, and how long does the checked result take today?
- What does a good result look like?
- Which input examples represent normal and unusual cases?
- Are we permitted to use those inputs in the proposed tool?
- What can our existing software already do?
- Who owns the process and has time to run a trial?
- Who reviews the output, and against what evidence?
- When will we stop the test, and how will manual work continue?
- How will we use any capacity released, and compare quality as well as time?
How to read your answers
Missing evidence gives you a preparation task. If the baseline is unknown, observe the work. If the inputs are unsuitable, resolve that before uploading them. If nobody can review the result, name a reviewer before choosing a tool.
Frequency is one factor, not the whole decision. Consider the effect of an error, the effort of checking it and how often the process changes. The checklist deliberately has no weighted score that could hide a serious gap.
What a paid assessment adds
Advice can help you challenge the proposed workflow, compare options and define a useful trial. Ask what evidence will be examined and what decision the work is meant to support. Do not assume an assessment includes implementation.
AI Consulting is A$1,195 per month + GST, with a call each fortnight, clarity on what to build in Claude or ChatGPT and practical guidance on how to do it. No lock-in. A build is separately scoped. Custom development is A$150/hour + GST, with a separate project scope, estimated hours and terms before work starts.
Gaps to look for during an assessment
Look for missing source information, an unclear owner, review work left out of the estimate or an existing feature that has not been evaluated. These are possible findings to investigate, not claims about how often they occur.
Ask the people doing the work to try the proposed approach. Check where they hesitate, what they correct and whether it affects a customer or another team. Their feedback belongs in the decision to continue.
What to do after the assessment
Choose one preparation task or a small supervised trial. Name its owner and agree a review point. Keep intended savings separate from measured results. What to automate first helps compare candidates, while AI adoption consulting covers involving the team.
Official resources
Explore the tools and check current details with the people who make them.
