Why Most Small Businesses Are Overthinking AI
The gap between AI hype and what a tradie or restaurant owner actually needs. Pick one painful task and automate it this week.
A restaurant owner asked me last month if he should build a custom AI model trained on his menu data. He'd read an article about fine-tuning and thought that was the way forward.
His actual problem? He was spending two hours every Monday manually writing the weekly specials email.
He didn't need a custom model. He needed a ChatGPT prompt and 15 minutes of setup.
This is what I see constantly. Small business owners hear the AI hype — custom models, agents, automation pipelines — and think they need something complex and expensive. They don't. They need the boring, practical version of AI that solves one specific problem.
The gap between what the tech industry is talking about and what a tradie, restaurant owner, or franchise operator actually needs is enormous. AI Twitter is debating AGI timelines. The plumber down the road just needs someone to answer his phone when he's on a job.
Here's my framework for thinking about AI as a small business owner: forget the technology entirely. Start with the pain.
What task do you or your staff do repeatedly that's tedious, time-consuming, and doesn't require deep expertise? That's your AI target. Not the complex strategic decisions. Not the creative work that requires industry knowledge. The repetitive stuff.
In my QSR world, here are real examples of where AI creates immediate value:
Shift notes. Every store manager writes shift notes after every shift. What happened, what went wrong, what needs attention. Most of them are terrible — either too vague to be useful or so long nobody reads them. AI can take bullet points and turn them into structured, consistent shift notes in seconds. Same format every time. Easy to scan. Easy to action.
Compliance tracking. Food safety logs, temperature checks, cleaning schedules, training records. All of this can be streamlined with AI-powered forms that flag anomalies, auto-complete based on patterns, and alert managers when something's overdue. Not a custom model. Just smart forms.
Roster analysis. Most QSR managers build rosters based on gut feel and last week's pattern. AI can analyse sales data by hour, day, and season, then recommend staffing levels that match actual demand. The savings on labour cost alone pay for any AI tool within a week.
Menu analysis. Which items are selling? Which have the best margin? Which combinations are popular? This data exists in the POS system, but nobody has time to analyse it regularly. A simple AI prompt that processes your weekly sales data and highlights opportunities takes five minutes to set up.
Customer complaint responses. Google reviews need responses. Social media comments need responses. Writing thoughtful, on-brand replies to every review is important but tedious. AI can draft responses that you review and send. Two minutes instead of twenty.
Notice what all of these have in common: none of them require a custom model. None of them require a data scientist. None of them cost more than a basic AI subscription.
For trades businesses, the opportunities are just as clear.
Quote follow-ups. How many quotes go out and never get followed up? AI can draft personalised follow-up emails based on the original quote details. "Hi Sarah, just checking in on the bathroom renovation quote I sent through last Tuesday. Happy to answer any questions." Sent automatically three days after the quote.
Job scheduling. AI can optimise daily routes based on job locations, estimated duration, and travel time. Not a complex logistics system — just common sense applied to a map.
Invoice reminders. Instead of chasing payments manually, AI can send personalised, escalating reminders. Friendly at first, firmer if they're overdue. The tone adjusts automatically based on how late the payment is.
The pattern is always the same: identify the repetitive task, find the AI tool that handles it, set it up in an afternoon, and move on. You don't need to understand transformers or attention mechanisms. You need to understand your own business well enough to spot the bottlenecks.
Here's what I'd recommend if you're a small business owner who hasn't started with AI yet:
This week, pick one task. Just one. The most annoying, repetitive thing you do. Write it down. Describe exactly what the inputs are, what the output should look like, and how often you do it.
Then try doing it with ChatGPT, Claude, or whatever AI tool you have access to. Copy in your inputs, tell it what you want, and see what comes out. It won't be perfect the first time. Refine the prompt. Try again.
If the output is 80% as good as what you'd produce manually, congratulations — you just saved yourself hours per week. Polish the last 20% by hand. That's still a massive win.
Don't build a custom model. Don't hire an AI consultant. Don't buy an enterprise platform. Just start with one task and a basic tool.
The businesses that win with AI won't be the ones with the most sophisticated technology. They'll be the ones that identified their real bottlenecks and automated them first.
If you want to bounce ideas about where AI fits in your business, send me an email — daine@dainereid.com. No pitch. Just happy to help you think through it.