Would AI actually help with this task? And if several tasks could benefit, which should you work on first?
I designed and built two tools to help answer those questions. The AI Suitability Matrix helps you think through whether a task is a good candidate for AI. The Automation Priority Matrix helps you compare those candidates and choose where to begin, considering the benefit, the work needed to get started, and the people and safeguards involved.
AI interest can move faster than a clear business case. Teams need a consistent way to examine fit, compare opportunities, and expose barriers before building.
I translated practical AI strategy questions into interactive decision tools, defining their criteria, scoring rules, recommendations, and user experience.
A checklist that suggests where AI could help, plus a comparison tool that ranks tasks, shows benefit versus ease of getting started, and flags serious concerns.
Both tools are publicly available. They help people discuss options and plan next steps. Their scores do not predict financial returns or guarantee a successful implementation.
Answer a short checklist about the task you have in mind.
Users can add up to two process-specific criteria. Each Yes contributes one point; each No contributes zero. The maximum is eight, nine, or ten according to the number of custom criteria added.
These score ranges stay the same if you add your own questions. A score of four or more also opens a guided set of questions about your task and tools, leading to suggestions for how AI could support the work.
In this prototype, the question about ethical risks counts as one point. Answering No does not automatically stop the tool from recommending AI. That means a high score can still hide a serious concern. Check the individual answers and resolve questions about privacy, fairness, responsibility, and expert judgment before taking the task forward.
Compare tasks that you have already judged appropriate for AI.
You might have several good ideas: drafting weekly reports, organizing customer feedback, or preparing meeting notes. This tool helps you compare them using the same questions rather than choosing whichever sounds most exciting.
For each task, answer 12 questions on a scale from 1 to 5. A higher rating means a greater benefit or better conditions for getting started. The tool combines your answers into an overall priority score out of 100.
Some answers count more than others. The four areas below show how much each contributes to that overall score.
Would this save meaningful time, improve quality, reduce frustration, or help more people?
Are the steps clear, the information available, and the tools easy enough to connect? How soon could you test a useful version?
Will people use it? Is someone responsible for introducing it, gathering feedback, and maintaining it?
Can you protect the information involved, catch mistakes, and keep the right people in charge of decisions?
I gave the greatest weight to useful outcomes and practical delivery. A good idea also needs people who can use and maintain it, with safeguards appropriate to the work.
Two answers act as checkpoints, called risk gates. They prevent a promising score from hiding a problem that needs attention first.
A rating of 1 means serious barriers to using the information safely. For example, a task may need confidential employee data that is not approved for use in the proposed AI tool. Resolve that barrier before proceeding.
A rating of 1 means mistakes or the need for human judgment are difficult to manage. For example, an AI error could affect someone’s employment and be hard to detect or challenge. Keep the task human-led unless effective review and accountability can be designed.
If either answer is 1, the tool marks the task risk review required. Its score stays visible so you can see the potential benefit, but it cannot be called a Quick win. It also moves behind every task that has passed these two checkpoints. Changing the plan and resolving the concern may make it worth reassessing.
The overall priority score is out of 100. It combines all four areas and helps rank tasks that have passed the risk checkpoints.
The map answers a different question: how valuable is this task, and how ready are we to start? It uses two separate scores, each also out of 100: one for Value and one for Readiness. The tool uses 60 out of 100 as the dividing line between lower and higher scores on each side of the map. Sixty is a cutoff on a 100-point scale, not a different scoring scale.
For example, Value of 80/100 and Readiness of 40/100 suggests a useful opportunity that needs preparation: a Strategic bet. If Readiness rises to 70/100, it becomes a Quick win, provided neither risk checkpoint has stopped it.
Useful and ready to start: Value and Readiness are both at least 60/100, with no risk gate triggered.
Useful, but needs preparation: Value is at least 60/100 and Readiness is below 60/100.
Ready to start, but offers less benefit: Readiness is at least 60/100 and Value is below 60/100.
Both Value and Readiness are below 60/100. Improve the idea or address the barriers before investing more effort.
Each 1–5 rating earns a share of the points available for that question: 1 earns none, 3 earns half, and 5 earns all. The tool adds those points and rounds the result to produce the overall score out of 100.
| Dimension | Criterion | Weight |
|---|---|---|
| Value & impact | How meaningful would the outcome be? | 15% |
| Value & impact | How much friction exists today? | 10% |
| Value & impact | How often does this process happen? | 10% |
| Value & impact | How many people or decisions would benefit? | 5% |
| Readiness & speed | How clear and repeatable is the process? | 8% |
| Readiness & speed | How ready are the information and examples? | 8% |
| Readiness & speed | How easy is it to connect the tools involved? | 7% |
| Readiness & speed | How quickly could a useful first version ship? | 7% |
| Adoption & ownership | How naturally would this fit into people’s work? | 8% |
| Adoption & ownership | How strong are ownership and change readiness? | 7% |
| Responsible use | How manageable are privacy and security concerns? | 7% |
| Responsible use | How manageable are errors and human judgment? | 8% |
The calculation for each question is (rating − 1) ÷ 4 × its weight. Each category score is calculated from the points earned within that category and expressed out of 100. The weights and the 60/100 dividing line are design choices for comparing options, not proof that an investment will succeed.
Save several tasks to compare their rankings and positions on the map. You can download the results as a CSV file for discussion or planning. Saved assessments remain in the browser you used to create them.
Imagine rating a weekly-report drafting task 5 on every question. It scores 100/100 and is labeled a Quick win. Those ratings still need evidence: approved information, useful drafts, a person to review them, and someone to maintain the workflow.
Now imagine the reports require confidential information that cannot be used safely in the proposed tool. Change the privacy/security answer to 1. The score is still 93/100, but the label becomes Promising—risk review required. The task moves behind those that have passed the risk checkpoints. Its potential value does not remove the need to resolve that concern.
Use the results to discuss what to try, what to prepare, and what to pause. Include the people who do the work and those responsible for reviewing it. Test a small version, check whether it helps, and update the ratings as you learn. These tools support that conversation; specialist advice may still be needed for legal, security, or other sensitive decisions.
The readiness assessment addresses a related question at a different level: what organizational conditions support responsible adoption?
Start with a real process. Examine fit, compare priorities, and identify what needs to be true before moving forward.