AI Use Case Prioritisation Matrix
Not sure which AI initiative to tackle first? Score up to five use cases across six strategic dimensions and get two numbers, not one: a business merit score and the autonomy clearance tier your governance readiness earns. A strong score on five dimensions no longer hides a dangerous one on the sixth.
Listen to this briefing
AI Use Case Prioritisation Matrix
Name each AI initiative you are considering. You start with three; add up to two more if you have them. Rate each one on the six dimensions below by picking the rung that matches your situation, not a raw number, so two different people scoring the same use case land on comparable results. Scores update in real time. Autonomy Readiness works differently from the other five: instead of averaging into the score, it sets the autonomy clearance tier the use case earns.
These five dimensions make up the Business Merit Score below. All five count equally by default. This scoring model is RICE-inspired, not a direct implementation of it: the confidence discount below works the same way RICE's Confidence term does, but the dimensions here are adapted for AI use case selection rather than RICE's original Reach, Impact, Confidence, and Effort. If Strategic Fit matters more to your organisation than Speed to Value, say so here, it re-weights every use case's merit score the same way. Autonomy Readiness is scored per use case too, but is deliberately left out of this weighting: a strong business merit score should never be able to average away a governance gap, so it is shown as a separate clearance tier instead.
This is a self-assessment built from the scores you enter, not an independent evaluation of these initiatives. It is not procurement, financial, or legal advice, and the recommended steps above are a starting point for your own due diligence, not a substitute for it. Your scores stay in your browser and are not sent to Terence Kok or reviewed by anyone.
What each dimension measures and why it matters.
Strategic Fit
Does this use case move a needle that your leadership cares about? A score of 1–3 means you are solving a problem nobody is measuring. A 9–10 means this directly enables a board-level priority. The highest-scoring use cases on this dimension are the ones that will get funded, resourced, and sustained past the pilot.
Data Readiness
AI without data is a framework without content. Score this dimension based on whether the data exists, is accessible, is clean, and is sufficient in volume for the model you intend to deploy. A score below 5 indicates data preparation work that will extend your timeline and cost significantly, as a directional pattern rather than a measured statistic, often by 3–6× the AI development cost itself.
Ease of Implementation
Technical complexity, integration requirements, change management load, and vendor dependency all affect this score. A 9–10 means an off-the-shelf solution with minimal integration. A 1–3 means custom model development, complex data pipelines, and significant organisational change, a category that requires a dedicated implementation team and realistic 12–18 month timelines.
Expected ROI
Consider both direct financial return (cost saved, revenue enabled) and indirect value (risk reduced, quality improved, capacity freed). Use your AI ROI Calculator to generate a defensible financial estimate before scoring this dimension. A 9–10 means a measurable, verifiable return within 12 months. A 1–3 means the value is theoretical, delayed, or difficult to attribute.
Speed to Value
How quickly can you demonstrate a result? This is not the same as ease of implementation: a use case can be technically simple but require 12 months of data collection before the model can make accurate predictions. Score 9–10 if you can deliver a measurable outcome in 4–8 weeks. Score 1–3 if the baseline measurement, training data collection, and model validation add up to over 12 months before any business value is visible.
Autonomy Readiness
A high-impact use case that also acts with meaningful autonomy, taking actions rather than just producing a recommendation, carries governance requirements a purely advisory tool does not. Score this dimension on whether your identity, logging, human-checkpoint, and audit controls are built for the level of autonomous action this use case will exercise, not on the use case's potential value. Unlike the other five dimensions, this score is not averaged into the Business Merit Score above. It sets the autonomy clearance tier instead: 1–3 clears the use case for checks and reports only, 4–6 clears it to draft an action and wait for a person to approve it, and 7–10 clears it to check and act entirely on its own, per the OWASP Agentic AI governance work and the Cloud Security Alliance's autonomy-tier guidance. A high score elsewhere can no longer outrank a use case ahead of the governance work it still needs, because the two numbers are never blended together.
I built this after watching too many leadership teams argue about which AI project to start with using nothing but instinct and whoever spoke loudest in the room. Score up to five use cases against the same six dimensions, strategic fit, data readiness, ease, ROI, speed to value, autonomy readiness, and the ranking does the arguing for you. I keep autonomy readiness out of the blended score on purpose: it earns the project a clearance tier instead, so a strong idea with a governance gap can't average its way past that gap. It won't tell you your idea is bad. It will tell you which one is ready to move first, and at what level of independence, which is usually not what everyone assumed.
