AI tools have never been more capable or more affordable. The businesses getting the most from them have a few things in common: a clear first use case, tidy data, systems the AI can reach, and simple guidelines that give everyone confidence.
The good news is that none of this needs a big budget to start. Here's the checklist we use to help clients get AI-ready, plus the advice we give leaders along the way.
Start with the work, not the tool
The best AI projects begin with a task your team already does, rather than a product someone saw in a demo. Look for work that ticks three boxes:
- Repetitive: it happens many times a week, in much the same way.
- Text or data heavy: it's mostly reading, writing, sorting or summarising.
- Easy to check: a person can confirm the result is right in seconds.
Tasks that tick all three are your best first candidates. Before you change anything, note how long each one takes today, so you can measure the improvement later.
Where AI fits best
These are the use cases we see deliver early wins across almost every industry.
| Use case | What AI does | A good first step |
|---|---|---|
| Customer enquiries | Drafts replies and routes each enquiry to the right person | Start in draft mode, with a team member approving each reply |
| Internal knowledge | Answers staff questions from your policies, procedures and manuals | Start with one well-maintained set of documents |
| Documents and forms | Pulls key details from invoices, applications and reports | Pilot on one document type with clearly defined fields |
| Reporting | Summarises your data and highlights trends in plain English | Connect it read-only to one trusted data source |
| Quotes and proposals | Drafts quotes and proposals from your own templates | Use your approved templates and pricing rules |
The AI readiness checklist
Work through these six areas in order. Each one makes the next easier.
1. Use cases
- Shortlist three tasks that are repetitive, text or data heavy, and easy to check.
- Choose one to pilot first, ideally where the time saved will be easy to see.
- Record how long the task takes today and how often it happens.
- Name a business owner for the pilot, not just a technical lead.
2. Data
- Find where the information for your chosen task lives: which systems, folders and documents.
- Tidy it up: retire outdated versions, merge duplicates and fill obvious gaps.
- Identify any personal, financial or commercially sensitive information.
- Decide exactly what the AI may read, and what stays out of scope.
3. Systems and connections
- Choose an approved AI assistant for the business, so staff know which tool to use.
- Check which of your systems have APIs the AI can connect to.
- Connect the AI to live data through secure connectors, such as MCP connectors, so answers come from your systems.
- Start with read-only access, then add actions once the team is comfortable.
4. Guidelines
- Write a one-page guide covering what to use AI for and what information to keep out.
- Keep a person reviewing AI output in the early stages, especially anything customer-facing.
- Check your AI provider's data terms, including where data is stored and whether it's used to train their models.
- Review the plan against your privacy obligations and any industry rules you follow.
5. Pilot
- Run a short pilot of a few weeks with a small, enthusiastic group.
- Measure the results against the baseline you recorded.
- Gather feedback and refine the instructions, templates and process as you go.
- Share what you learn with the wider team, including the quick wins.
6. Scale
- Roll out what works, with short training sessions and your one-page guide.
- Pick the next use case from your shortlist.
- Review usage, results and guidelines every quarter as the tools improve.
Advice for leaders
- Try the tools yourself. Half an hour with an AI assistant on your own inbox or reports will sharpen your instincts for where it fits.
- Your people are the multiplier. They know the tasks best. Involve them early and position AI as taking on the busywork so they can focus on the work they enjoy.
- Start small, scale fast. One well-run pilot builds more confidence than a big program, and the second use case moves much faster than the first.
- Data is the foundation. Every hour spent tidying and connecting your data pays off in every AI project that follows.
- Choose flexibility. Open standards like MCP let you switch or add AI assistants later without rebuilding your connections.
How James Anthony Consulting can help
James Anthony Consulting helps organisations put AI to work with confidence. We help you pick the best first use cases, connect AI assistants to your systems with secure custom MCP connectors, and build practical AI tools on platforms like AWS Bedrock and Azure AI Foundry.
A short discovery workshop is a great place to start. You'll leave with a shortlist of use cases, a clear view of your data and systems, and a plan for your first pilot. If you're ready to explore what AI can do for your business, get in touch for a chat.
