AI at Work10–15 minutes

AI Use Case Canvas

Find the right problem before choosing AI.

Runs in your browser

Use this when: deciding whether an existing business problem is a good AI opportunity before planning a project.

WHAT YOU'LL ACHIEVE

Assess whether a business problem is a sensible candidate for AI, what success would look like and what needs to be in place before starting.

THE TOOL: AI USE CASE CANVAS

Start with the work as it is.

WORKSHEET 01

Open the Quick Opportunity Check

QUICK OPPORTUNITY CHECK

Could this workflow be automated?

A conservative, rule-based screen. Choose categories only, never enter sensitive information here.

How often does this task happen?How repeatable is the task?How structured are the inputs?How stable is the process?How often are there exceptions?What information sensitivity applies?What could happen if the result is wrong?How much human judgement is needed?

Complete all eight choices to see the screening rationale.

01Business problemWhat is happening today and why is it a problem?

02Current processHow is the work done now?

03Desired outcomeWhat should improve?

04AI's proposed roleAssist, recommend, classify, generate or take an authorised action?

05Information requiredWhat data, documents or systems are needed?

06Risks and controlsWhat could go wrong and what safeguards are needed?

07Success measuresHow will value and quality be measured?

08Owner and next stepWho is responsible and what should be tested first?

YOUR CANVAS / browser-only

WORKED EXAMPLE

Repetitive policy enquiries

An HR team receives a high volume of repetitive policy enquiries.

The proposed use case is an AI assistant that retrieves relevant approved policy information and drafts answers. Success could be measured through response accuracy, time saved, employee satisfaction and the number of enquiries needing escalation.

Before proceeding, the team would need to confirm document quality, access permissions, ownership and the consequences of incorrect answers.

WHAT TO CHECK

AI isn't automatically the improvement.

Don't assume that introducing AI is automatically better than improving the existing process or using simpler automation.