68% of Small Businesses Are "Using AI." Most Are Doing It Wrong.

There's a stat doing the rounds: 68% of small businesses are using AI.

Sounds like the revolution has arrived. Nearly 7 in 10 businesses have adopted AI.

Except they haven't.

A March 2026 DigitalApplied study found that the vast majority of that 68% are at "Tourist" level. They've asked ChatGPT to draft an email. Played with an image generator. Used AI to write a social post.

That's not adoption. That's curiosity. And curiosity doesn't build competitive advantage.

The Five Stages of AI Adoption

At FarFront, we assess every business against five stages — because the jump from each one to the next requires different things.

Stage 1: Unaware. You've heard of AI. It's not part of how you run your business. Roughly 30% of trades businesses are here — not because they're behind, but because they're busy.

Stage 2: Exploring. You use ChatGPT for emails or proposals. You know there's potential. Nothing in your business has actually changed. This is the "68%" — the stat that sounds impressive but means almost nothing.

Stage 3: Experimenting. You've picked 1-3 specific processes and you're testing AI on them systematically. Automated enquiry response. AI-assisted quoting. Invoice follow-up sequences. You're measuring what works. About 10-12% of small businesses are here.

Stage 4: Implementing. AI is embedded across most core processes. Humans manage exceptions and make decisions. AI handles the routine. The business operates at a level that would typically require 2-3x the admin staff. Less than 2% of businesses are here.

Stage 5: Optimising. The systems measure their own performance and suggest improvements. The business learns continuously. The gap to competitors widens every month without additional effort. Almost nobody is here yet — but the businesses that reach Stage 4 first will get here first.

Why 2026 Is the Tipping Point

AI has been "coming" for years. Three things changed in 2026 — all at once.

The cost line crossed. AI costs dropped over 90% in 12 months. The tools that connect AI to your job management software shipped native integrations in 2025. You don't need a developer anymore. For the first time, the economics work for a 15-person plumbing business.

Customer expectations shifted permanently. Some trades businesses ARE responding to enquiries in 5 minutes now. Customers don't know it's AI — they just know that plumber replied instantly and you didn't. The bar has moved. If you're still at 24-48 hours, you're not slow — you're invisible.

This is the window where small beats big. A 15-person electrical contractor can implement AI in 4 weeks. A 200-person operation takes 6-12 months — procurement committees, IT approvals, change management. Right now, you can move faster than companies with 10x your resources.

That window is temporary. The big players will catch up. They always do. But right now — in 2026 — most large trades and construction companies are still evaluating. Still running pilots. Still waiting.

This is the 12-18 month window where a nimble business can build systems, accumulate data, and establish a customer experience advantage that compounds every month. By the time larger competitors arrive, the early movers have 18 months of compounded improvement baked in.

That's not a lead. That's a moat.

The Mistake: Automating What You Already Have

Here's where most AI advice gets it wrong.

The standard playbook: take your existing processes and add AI. Make them faster.

The problem: most trades business processes weren't designed — they evolved. Shaped by whatever software was available, whatever the previous admin person set up, whatever worked well enough when there were 5 employees.

Bolting AI onto a broken process doesn't fix it. It just breaks faster.

If your quoting process has 8 steps because that's how your job management software works, adding AI to step 4 makes step 4 faster. You still have 8 steps. You've improved efficiency. You haven't changed the game.

The better question: If you designed this process from scratch today — knowing what AI can do — what would it look like?

The answer is almost never "the same 8 steps, but faster."

This is the distinction between Improve and Redesign — and it's the decision that determines whether you end up at Stage 3 or Stage 4.

The technology for the right column exists today. The gap isn't capability — it's that nobody has redesigned the processes. They're running 2015 workflows with 2026 tools.

The Redesign path requires three things simultaneously: trades domain knowledge, process design expertise, and AI technical knowledge. Almost nobody has all three. That's why most advice defaults to "automate what you have" — it's the easy answer, not the right one.

What the Gap Actually Looks Like

Same customer. Same job. One business at Stage 4, one at Stage 2.

The Stage 4 plumber responds in 5 minutes. The Stage 2 plumber responds in 2 days. Job gone.

The Stage 4 electrician quotes on-site from a photo. The Stage 2 electrician quotes from the office the next day. Revenue gone.

The Stage 4 builder's compliance docs write themselves. The Stage 2 builder's admin spends 6 hours a week on SWMS. Margin gone.

The Stage 4 business invoices on completion with AI-tuned follow-up. The Stage 2 business invoices next week. Cash flow gone.

Multiply that across 50 enquiries a month, 12 months a year. The Stage 4 business isn't just winning more work — it's winning the best work, building the strongest reviews, attracting the best staff, and reinvesting the margin into the next improvement. Every month, the gap widens.

Why Most Businesses Stay Stuck

Complexity, not cost. The Bookipi March 2026 survey confirmed it: 500 AI tools, no clear guidance, nobody explaining what "good" looks like for a 15-person plumbing business. So people don't start.

The "I'll get to it" trap. Every trades owner says: "I know I should be doing more with AI. I just haven't had time." They're right — because they're buried in the exact operational chaos that better systems would fix.

No one shows them the path. Not which tool to buy. The path: which processes to change, in what order, and whether to Improve or Redesign each one. Without that, AI adoption is just shopping.

What to Do About It

Moving from Stage 2 to Stage 3 takes about 15 hours of focused setup over 4 weeks. Pick 2-3 high-impact processes. Implement properly — not just try a tool, but embed it into how work gets done.

For each process, make the call: Improve or Redesign?

The businesses that only Improve will reach Stage 3 — a genuine competitive advantage.

The businesses that Redesign the right processes will reach Stage 4 — a structural moat that compounds every month and becomes very hard to reverse.

68% of your competitors think they're "using AI." They're not. The window is open. It won't stay open.