Every week I talk to a founder who feels behind on AI. They’ve read the headlines. They’ve seen competitors claim they’re using it. They know they should be doing something, but they don’t know where to start — and they’re afraid of looking foolish by asking basic questions.
Here’s the truth: most of those competitors claiming to use AI are doing one of three things. They bought a tool that has “AI” in the marketing copy. They gave everyone ChatGPT accounts with no strategy. Or they’re bluffing entirely. Genuine, strategic AI adoption at the SME level is still rare — which means there’s a real advantage for founders who get it right now.
This guide is specifically for founders who don’t code. No jargon, no acronyms without explanation, no assumption that you know what a large language model is. Just a practical framework you can apply this week.
Step 1: Find your highest-volume repetitive tasks
AI delivers the best ROI on tasks that are repetitive, high-volume, and currently done by humans who could be doing something more valuable. Don’t start by thinking about AI capabilities. Start by thinking about where your team wastes time.
Ask every team lead this question: what task does your team do repeatedly that follows the same pattern every time? The answers are usually some combination of responding to similar customer enquiries, extracting data from documents and entering it into systems, generating reports from the same data sources, qualifying leads by checking the same criteria, and formatting or reformatting content across platforms.
A home services company I work with in the Thames Valley discovered that their office manager spent 15 hours per week copying information from customer enquiry forms into their scheduling system, then sending confirmation emails. That’s one person’s time, two full days a week, on a task that follows the exact same pattern every time. That was their AI starting point — not because AI is exciting, but because automating that process freed up 15 hours of human capacity every week.
Step 2: Evaluate your AI readiness honestly
Not every business is ready to implement AI. Readiness isn’t about technical sophistication — it’s about whether your business has the foundations that AI needs to work.
There are three foundations that matter. First, do you have your data in digital form? AI can’t work with sticky notes and filing cabinets. If your customer records, processes, and knowledge are locked in people’s heads or paper files, the first step is digitisation, not AI. Second, do you have a clear process? AI automates processes that exist. If your team handles every situation differently with no standard approach, AI will automate chaos. Document the process first. Third, do you have someone who can own it? Every AI initiative needs a business champion — someone who defines success, tests the output, and pushes for adoption. Technology alone doesn’t change anything.
The AI Readiness Quiz I built covers all three dimensions and gives you a personalised score with specific recommendations. It takes five minutes and it’s the simplest way to understand where you stand.
Step 3: Start with one pilot — not a strategy document
The biggest mistake I see founders make is trying to create a comprehensive AI strategy before they’ve run a single experiment. You wouldn’t write a five-year marketing plan without ever running an advert. Don’t create an AI roadmap without ever deploying an AI tool.
Pick the single highest-impact task from Step 1. Set a clear goal: reduce the time spent on this task by 50% within 30 days. Choose the simplest tool that could work. Run it for a month. Measure the results. That one experiment will teach you more about AI in your business than any strategy workshop.
Here’s what good pilots look like at the SME level. Customer enquiry auto-response: use an AI chatbot to handle the 60 to 70% of enquiries that are repetitive (opening hours, pricing, availability), escalating the rest to humans. Document processing: use AI to extract data from invoices, contracts, or application forms and populate your systems automatically. Content generation: use AI to draft first versions of blog posts, social media content, or proposal templates that your team then reviews and customises.
Step 4: Measure ROI in hours, not hype
AI vendors will sell you on transformation. Ignore that word. Measure AI success the same way you measure any business investment: what did it cost, and what did it save?
The simplest ROI calculation for an AI pilot is this. Count the hours your team currently spends on the target task per week. Multiply by the loaded cost per hour (salary plus NI plus benefits divided by working hours). That’s your current cost. After implementing AI, count the hours again. The difference multiplied by 52 is your annual saving. Compare that to the implementation cost plus ongoing subscription.
The home services company I mentioned earlier was spending roughly £18,000 per year on the data entry task (15 hours per week at approximately £23 per hour loaded cost). The AI automation cost £6,000 to implement and £150 per month to run. First-year ROI was positive by month five. Every subsequent year saves the full £18,000 minus the £1,800 running cost. That’s the kind of maths that makes AI adoption a straightforward business decision, not a leap of faith.
How to spot AI washing
AI washing is the practice of slapping “AI-powered” onto products that are doing little more than basic automation, rules-based logic, or simple database lookups. It’s everywhere, and it wastes money.
Three red flags to watch for. First, the vendor can’t explain what the AI actually does in plain language. If they hide behind jargon and refuse to show you a working demo with realistic data, be sceptical. Second, the promised accuracy is 99% or higher. Real-world AI is messy. For most business tasks, 85 to 95% accuracy is excellent. Anyone promising near-perfect accuracy is either measuring wrong or lying. Third, there’s no human review step. Any AI system that touches your customers, your finances, or your compliance should have a human reviewing critical outputs. If the vendor says their AI needs no oversight, they don’t understand the risks.
The founder’s AI decision framework
When evaluating any AI opportunity, ask yourself these four questions. Is this task repetitive enough that a machine could learn the pattern? Is the cost of doing it manually high enough to justify automation? Can I measure the outcome clearly? Am I comfortable with 85 to 95% accuracy, with a human catching the rest?
If all four answers are yes, you have a strong AI candidate. If any answer is no, either fix that gap first or look for a different task. This framework has saved my clients from dozens of bad AI investments — and surfaced genuine opportunities they would have missed.
You don’t need to become technical. You need to become strategic about which problems are worth solving with AI and which aren’t. That’s a business skill, not a technology skill. And if you want someone to walk through this framework with you for your specific business, that’s exactly what a fractional CTO engagement is for.
Start with the AI Readiness Quiz. Five minutes, honest answers, and you’ll know exactly where you stand.