Is your business ready for AI?
Before choosing a model or buying another subscription, choose a job worth improving. This checklist helps you assemble the evidence needed for a useful first AI project.
A business is ready for a focused AI pilot when it has a defined problem, usable and permitted inputs, someone accountable for reviewing the result, and a way to measure success. Readiness is specific to the task; it is not a single score for the whole company.
Start with a task you can describe
“We need to use AI” is not a brief. “Our team spends time finding the current procedure for customer queries” is something you can investigate. Identify who does the work, what triggers it, what information they use and where the completed result goes.
Record a baseline before the pilot. Sample ordinary cases and awkward ones. A faster draft is not an improvement if the checking takes longer, so measure the whole task rather than the model’s response time alone.
- Name the process owner and the people who will use the result.
- Record task volume, handling time and common exceptions.
- Define an acceptable output and the errors that matter most.
Check the information and the right to use it
Gather representative inputs rather than preparing only clean demonstration examples. Check dates, duplicates, incomplete records and contradictory documents. For a knowledge assistant, decide which source wins when two procedures disagree.
Confirm how the system can access the information and which users may see it. A file that exists is not necessarily available through an API, covered by your licence or suitable for every employee. Record unanswered access questions as dependencies, not assumptions.
Decide what happens when the result is wrong
Choose the reviewer, the review interface and the actions that require approval. Include a route for missing evidence and unavailable systems. A pilot can begin with read-only access and draft outputs while the team learns what to expect.
Agree how to stop the process and return to manual work. If a task updates a record, design protection against repeated events. If information changes, decide who updates the source and how quickly the change needs to appear.
- Who checks the output before it affects a customer or a business record?
- What stops repeated attempts creating duplicate work?
- Who owns updates, support and the decision to pause the system?
Work through your first use case
Select only the statements you can support with evidence. The interactive checklist below runs in your browser and does not submit your answers. It counts preparation items; it does not certify readiness or predict return on investment.
An unchecked item is a question to resolve. Permission, review and recovery gaps deserve attention before live data or real actions enter the pilot, even if most other boxes are checked.
Turn the answers into a pilot brief
Write a short brief containing the task, sample inputs, expected outputs, permitted sources, reviewer, measures and stopping conditions. Set a review point at which you will continue, change the design or stop. This gives both your team and a supplier something concrete to assess.
A readiness assessment is useful when the checklist raises questions that need investigation across systems or teams. The goal is an evidence-based decision, not a higher score.