AI Automation Readiness Checklist

Before investing in AI, check that you have the right problem, the right data and a plan to measure results. This checklist helps you pick a first use case that is likely to succeed. For context, read AI Automation for Small and Medium Businesses.

1. Identify candidate tasks

• List tasks your team repeats daily or weekly.
• Estimate hours spent per week on each.
• Note which tasks involve reading, writing, sorting or summarising text.
• Note which tasks follow clear rules (these may not need AI at all).

Example candidates: answering admission enquiries, extracting invoice data, categorising support emails, drafting product descriptions.

2. Check the inputs

• Is the information already digital (emails, PDFs, forms)?
• Are documents reasonably consistent in layout and quality?
• Do you have at least 30–50 real examples to test with?
• Are there languages other than English involved?

3. Define the output

• What should the automation produce (an answer, a record, a draft, a label)?
• Where should the output go (CRM, ERP, email, spreadsheet)?
• Who reviews the output before it is used?

4. Evaluate risk

• What happens if the AI is wrong? Low, medium or high impact?
• Is personal or sensitive data involved?
• Are there legal or compliance requirements?
• Is human review required for every item, or only uncertain ones?

Start with low-risk tasks where mistakes are easy to catch.

5. Estimate value

• Hours saved per week if the task is 70–80% automated.
• Faster response times for customers.
• Reduced errors or rework.
• Estimated monthly AI usage cost.

6. Prepare your team

• Nominate an owner for the pilot.
• Explain to staff how AI will support, not replace, their judgement.
• Agree how feedback will be collected.

7. Plan the pilot

• Choose one use case.
• Define success measures (accuracy, time saved).
• Set a short pilot period (four to six weeks).
• Decide what result would justify a wider rollout.

Scoring

Give each candidate task a score of 1–5 for value, data readiness and low risk. Multiply them. Start with the highest score.

Next steps

Share your shortlist with our AI business automation team. We can test your examples and give an honest view of what AI can and cannot do for that task.