Seven questions to check your AI Readiness
1. Are you already using AI, including through everyday business software?
AI use is not limited to specialist systems or standalone tools. It may already be built into software used for bookkeeping, customer service, marketing, recruitment, cybersecurity or data analysis. The key issue is whether anyone has a clear view of where it is being used.
A quick internal check should cover which tools are in use, who is using them and what tasks they support. It should also look at whether staff are entering customer, employee or business information into those tools. You may uncover useful experiments that could be shared more widely. You may also find public tools being used without anyone knowing what happens to the information uploaded.
There is no need for a technical audit at this stage. A short conversation with each team can show where AI is already part of the working day and where clearer guidance may be needed.
2. Do you know which business problem you want AI to solve?
Buying an AI tool because everyone is talking about AI rarely ends well.
Start with a real business issue. It might be a task that takes too long, creates avoidable errors or slows down customer service. Common examples include processing routine enquiries, preparing standard documents, reviewing large volumes of information or producing an initial sales forecast.
Take a small firm that receives the same types of customer questions every week. Staff are spending too much time writing similar replies and customers are waiting longer than they should. That is the problem. A chatbot may be one option, but it should not be treated as the answer before the issue has been properly understood.
Be specific about the result you want. "Use AI in marketing" gives little direction. "Reduce the time spent preparing the first draft of our weekly customer email" gives you something that can be tested.
Sometimes AI will be the right answer. In other cases, a process change or better use of existing software may work just as well.
3. Do your employees have the skills and confidence to use AI?
Employees do not need to understand how an AI model is built. They do need to know how to use a tool properly, question a weak answer and recognise when human judgement matters. Confidence will vary across a business. People who are already experimenting may need clearer boundaries. Colleagues who have avoided AI may need a practical introduction linked to their own work.
Training works best when it focuses on real tasks rather than abstract concepts. Staff should know how to give clear instructions, check facts, protect confidential information and improve an output instead of accepting the first answer.
The people closest to a process usually know where the frustrations sit. Involving them before choosing a tool can prevent the business from paying for something that looks impressive but does not fit the job.
Used well, AI can reduce routine work while people remain responsible for judgement, quality and customer relationships.
4. Is your business data accurate, accessible and secure?
AI can work quickly with information. It cannot know that a spreadsheet contains outdated figures or that two teams record the same customer in different ways. Look at the information behind the task you want to improve. Is it current? Is it recorded consistently? Can the right people access it? Does anyone know which version is correct?
Security needs the same attention. Staff may not realise that pasting text into a public AI tool can mean sharing it with an external provider.
Take particular care with:
- customer and employee records
- financial information
- contracts and commercially sensitive documents
- passwords, security details and intellectual property
Check the tool's privacy terms and settings. Where the position is unclear, keep confidential information out of it.
Good data management may feel like background work, but it often decides whether an AI project becomes useful or unreliable.
5. Do you have clear guidance on responsible AI use?
A short set of practical rules can prevent a lot of confusion.
Staff should know which tools are approved, what information must not be entered and when AI-generated work needs checking. They should also know who remains responsible for the final document, recommendation or decision.
Accuracy is an obvious concern. AI tools can produce incorrect information in a confident tone. Bias may also affect outputs, particularly where a tool supports recruitment, customer decisions or other areas involving people.
For a small business, one or two pages may be enough at the outset. Use plain language and examples drawn from the work people actually do. Update the guidance when new issues come up.
One point should remain clear throughout: a named person is responsible for anything the business sends, publishes or acts upon.
6. Can you measure whether AI is improving the business?
A tool can feel helpful without delivering much value.
Decide what you will measure before a trial begins. The right measure depends on the problem, but it could include:
- time saved
- fewer mistakes
- faster response times
- reduced costs
- better customer feedback
- more sales enquiries
- improved staff productivity.
Capture the current position. If preparing a report takes four hours today, record that. Test the new approach for a sensible period and compare the result.
Remember the less visible costs. Staff need time to learn, check outputs and fix problems. A monthly subscription may look inexpensive until several people need licences or the business needs outside support.
Small improvements can matter when a task is repeated often. Saving ten minutes once a month will not change much. Saving ten minutes fifty times a week may be worth pursuing.
Not every trial will work. A controlled test lets the business find that out before making a larger commitment.
7. Do you know where to access trusted advice and support?
The AI market is crowded. Product claims can be difficult to compare and the right choice for a large company may be unsuitable for a small firm. Good advice should begin with the business problem rather than a product demonstration.
Before making a significant investment, ask what the tool will improve, how data is handled, what training is included and how performance will be measured. You should also know what happens if the tool does not perform as expected. Independent guidance, sector examples and conversations with other SMEs can help cut through the noise. A business that tested a tool and decided not to proceed may have as much to teach as one promoting a successful project.
