Where to start with AI: a 30-day plan for North Wales business owners
A four-week plan that takes you from vaguely curious to using one AI tool that saves you measurable time, every week, with zero training spend.
The hardest question with AI is not what it can do. It is where to start. Most business owners I talk to have either tried it once and moved on, or done nothing because the noise is overwhelming.
This is a four-week plan that takes you from "vaguely curious" to "using one AI tool that saves you measurable time, every week". It is designed for business owners who want to act, not for IT teams running pilots.
Week 1: Find your time-drains
Forget the technology for a week. Open a notebook or a spreadsheet. Every time you do something this week that takes more than fifteen minutes and feels repetitive, write it down. Note how long it took.
By Friday you will have a list. It might look like this:
- Replying to enquiries (90 minutes a day, on average)
- Writing the weekly newsletter (3 hours)
- Responding to Google and TripAdvisor reviews (45 minutes, twice a week)
- Drafting quotes and proposals (4 hours)
- Updating the website with new dates, prices or posts (1 hour)
Pick the one that scores highest on two things: most hours, and most language-heavy. AI is best at language work. It is far less useful for jobs that involve physical doing or specialist judgment.
Week 2: Pick one task. Try one tool.
Now you have your task, pick one AI tool. Use ChatGPT or Claude on the free tier. The choice matters less than the commitment to actually use it. We have a full comparison if you want to think it through.
Spend the week using it ONLY for that one task. Resist the temptation to scroll through the feature list and try everything. The point is to get one thing working well, not to feel like you have explored AI.
Each time you use it, write a short note: how long did it take, was the output usable, what did you have to change. By Friday you will have a feel for whether AI is going to take twenty minutes a day off this task, two hours a week, or barely move it.
Week 3: Make it permanent
If week 2 worked, this is the week you make it stick. There are three things to do.
First, save your prompts. Whatever phrasing worked best, write it down somewhere you can copy-paste. Most people lose 80% of their AI productivity by retyping prompts every time. Build a saved-prompts file and refer back to it.
Second, share it with one other person. If you have employees, train one person on the same task. If you are a solo business owner, tell another business owner what worked. The act of explaining it sharpens your own understanding.
Third, decide on a paid tier or stay on free. The paid tiers - ChatGPT Plus, Claude Pro - cost £18 to £20 per month. If you are saving more than two hours a week, the maths is straightforward.
Week 4: Add one more
Only now do you add a second task. Look back at your week-1 list and pick the next-best one. Repeat the same pattern: try AI on it, see if it works, save the prompt.
The reason this matters: most AI rollouts fail because someone tries to introduce it to fifteen tasks at once and gets buried. The pattern that works is one task, then one more, then one more. Each one becomes a habit before the next is added.
What this looks like for different sectors
In tourism and hospitality, week 1 almost always points to guest communication or review responses. Both are pure language work. Both eat hours that could be spent welcoming guests properly. The five tourism prompts post has practical templates.
In professional services - accountants, solicitors, surveyors - the answer is usually document analysis. Reading sets of accounts, summarising long contracts, drafting reports. AI document processing is the deeper guide.
In construction and trades, the answer is often quotes and proposals. The actual building is not getting done by AI, but the paperwork that surrounds it can be.
In manufacturing and engineering, look at supplier comms, technical documentation, compliance write-ups.
In retail and e-commerce, look at product descriptions, social content and customer enquiries.
Avoiding the 'shiny object' trap
The biggest trap in AI adoption is collecting tools. There are AI products for everything: AI for note-taking, AI for graphic design, AI for video editing, AI for SEO, AI for sales, AI for forecasting. Most of them are good. None of them are useful unless you have made the time to integrate them into a real workflow.
The plan above only commits you to one tool, used for one task at a time. That is on purpose. The compounding return on getting one thing right is much greater than the return on having ten things half-installed.
The honest scope
This plan will not make you an "AI-driven business" in thirty days. That is a phrase from LinkedIn, not from real life. What it will do is give you one workflow that uses AI well, by the end of the month, with zero training spend and a maximum of £20 in software costs.
For most small businesses in North Wales, that is enough to free up a few hours a week. A few hours a week, repeated weekly, is what compounds into the genuine "I cannot imagine running this without it" moment, usually about six months in.
If you would like to walk through your week-1 list with someone who has done it many times, that is what a free 30-minute discovery call is for. The AI Breakfast Club training also runs through this kind of structured adoption in its first session.