Why Most Small Businesses Using AI Still Aren't Seeing the Results They Should

A Goldman Sachs survey published this year found that 93% of small business owners say AI has had a positive impact on their operations. That sounds like the story is over. AI works, businesses are happy, problem solved.

But look at the next stat from the same survey: 73% of those same owners say their business would benefit from more training and guidance to get proper value from AI.

So the majority of small businesses using AI are doing so without fully understanding how to get the most out of it. They're picking up fragments of value, but leaving most of the benefit on the table. That's not a technology problem. It's a knowledge problem, and it's more common than most people admit.

‍ ‍

The tool isn't the issue

When a small business owner says AI hasn't delivered what they expected, the instinct is to blame the tool. Wrong platform, wrong software, maybe they chose the wrong thing.

That's almost never the actual problem.

A study by INSEAD and Harvard tracked 515 companies that had access to the same AI tools, training budget, and resources. Half of them also received help figuring out where to apply AI in their specific workflows. The results were stark: those companies achieved 1.9 times higher revenue, found 44% more ways to use AI across their business, and were 18% more likely to convert new customers.

Same tools. Completely different outcomes.

The difference wasn't the technology. It was knowing exactly where to point it.

The question most business owners skip

The most common way a small business approaches AI is to start with the tool. They hear about something, sign up for a trial, and try to fit it around whatever they're already doing. Sometimes it sticks. More often, it quietly gets abandoned three months in.

The better starting point is a different question entirely: which specific task in your business takes the most time, follows the same steps every time, and doesn't actually require a human's judgment to complete?

That task, whatever it is, is your first automation.

It might be manually replying to the same enquiry emails every day. It might be chasing unpaid invoices. It might be updating a spreadsheet with information that already exists in another system. It might be drafting social posts for next week.

The businesses that see fast, measurable results from AI aren't the ones with the most sophisticated tools. They're the ones who identified a specific, repeatable problem and solved it. Then they found the next one.

Why the "where to start" problem is harder than it looks

A separate study found that 39% of UK businesses say the biggest barrier to AI adoption is simply not knowing where to use it. Nearly four in ten. And this isn't a lack of ambition; these are businesses that want to adopt AI but can't figure out the right entry point.

Part of the problem is the way AI is sold to small businesses. The marketing tends to focus on everything AI could do: write your content, manage your customer service, run your social media, automate your entire back office, which sounds impressive but gives you very little to work with on a Monday morning when you've got client work to deliver.

The practical reality is that the biggest, fastest wins from AI tend to come from the unglamorous stuff. Phone calls going to voicemail that shouldn't be. Customer enquiries sitting unanswered for hours. Manual data entry that takes someone a morning every week. These aren't the headline use cases you see in press releases, but they're the ones that free up real time and deliver real return quickly.

What "not seeing results" actually looks like

It's worth being specific about what underperforming AI adoption looks like in practice, because it doesn't always look like failure.

Often it looks like a business that has signed up for three or four AI tools, uses each of them occasionally, and would struggle to say exactly what they've saved in time or money. The tools work fine, they just haven't been integrated into any proper workflow. They sit alongside the existing process rather than replacing it.

Another common pattern is using AI for the visible stuff, generating a few social media posts, maybe writing a draft email, but not touching the operational tasks where the real time is being lost. It feels productive. The savings are minimal.

Neither of these is a disaster. But neither is extracting the kind of value the technology is capable of delivering.

Where to go from here

If you're using AI tools and not entirely sure whether they're making a material difference, the starting point is straightforward: map out where your time actually goes in a typical week. Be honest about it. Not what you'd like to be spending time on, but what you're actually spending time on.

Look for the tasks that are repetitive, predictable, and time-consuming. Those are the candidates for automation. Pick one. Sort it properly. Measure the time saved. Then move to the next one.

It's not a complicated framework. But it's the approach that the businesses genuinely getting value from AI are using, and it's a long way from buying a tool and hoping for the best.

We help businesses cut through the noise and build digital marketing that actually works. If you're not sure where AI fits into your marketing, get in touch, and we'll give you a straight answer.