A recent Dell study reported by IT Brief UK found that UK SME AI front-runners, the small and medium businesses furthest along with artificial intelligence, are saving noticeably more time than the firms still dabbling. The short answer to “why” is not that they bought cleverer software. It is that they picked a few painful, repetitive jobs, put the right tool on each one, and made it a habit rather than a one-off experiment.
If you run a small business and you are wondering whether the AI hype is real or just something for big corporates with data teams, the useful takeaway is this: the gap between the front-runners and everyone else is mostly about method, not budget, and free tools like the ChatGPT small business agent lower the barrier further. Below we break down what these firms do differently, which tools UK owners actually shortlist, what they cost in pounds, and how to copy the approach without hiring a single specialist.
What the Dell research actually says
Dell (see Dell UK) runs regular research on how businesses adopt technology, and the theme in this latest reporting is consistent with its wider “Innovation Index” work: a leading group of companies is pulling ahead on measurable outcomes, and time saved is one of the clearest. The study frames these leaders as AI “front-runners” who report stronger productivity gains than slower adopters.
We are deliberately not quoting a precise percentage here, because the headline figures vary between summaries and you should treat any single number with caution. The pattern, though, is the point worth acting on: businesses that treat AI as an operating habit, the same conclusion reached in Xero’s research on AI adoption and profitability get more back than those that treat it as a gadget. The same story shows up in Xero research on AI doubling accounting-firm profitability, and in free tools like the ChatGPT small business agent. You can read the original write-up on IT Brief UK for the framing, then judge for yourself against your own diary, a pattern also visible in Xero’s research on AI adoption profitability.
What front-runners do differently
Across this study and most credible AI-adoption research, including Xero research linking AI adoption to doubled profitability, the same handful of behaviours keep appearing. None of them require deep pockets.
- They start with a task, not a tool. Instead of “we should use AI”, they say “quoting takes us four hours a week, let’s cut that”. A specific job with a measurable time cost gives you something to test against.
- They pick jobs that are frequent and boring. Drafting emails, summarising meetings, first-draft proposals, chasing invoices, tagging support tickets. High volume plus low judgement is the sweet spot.
- They keep a human in the loop. The front-runners use AI to produce a first draft or a shortlist, then a person checks it. That is what keeps the time saving without the reputational risk.
- They connect tools together. The biggest gains come when the output of one system flows into the next without someone copying and pasting. This is where automation platforms earn their keep.
- They measure the before and after. Even a rough note of “this used to take an afternoon, now it takes 40 minutes” is enough to decide whether to keep going.
If you want a structured way to spot these jobs, our guide to automating repetitive tasks in a small business walks through how to audit a week and score each task for AI suitability.
The tools UK SMEs actually shortlist
Naming real products matters here, because “get an AI assistant” is useless advice. Below are the options a UK owner would genuinely consider, including a few smaller names worth knowing beyond the American giants.
For general assistants and drafting, the obvious three are Microsoft Copilot (strong if you already live in Word, Excel and Outlook), ChatGPT from OpenAI, and Google Gemini if your business runs on Google Workspace. A quieter option worth a look is Claude from Anthropic, which many owners prefer for longer documents and careful writing.
For the automation “plumbing” that connects your apps, the names to know are Zapier, the more visual and often cheaper Make, and for the technically confident, the open-source n8n, which you can self-host to control costs and data.
Beyond the household names, two specialists UK firms often overlook: Synthesia, a London-founded company that turns a script into a presenter-led video (handy for training or product explainers without a film crew), and Perplexity, a research assistant that cites its sources, which reduces the “the AI made it up” risk when you need real references. For customer-facing chat and support, Intercom has leaned heavily into AI answering, and it is a realistic shortlist entry if you handle a lot of repeat queries.
How the main options compare
| Tool | Best for | Indicative UK price | Watch out for |
|---|---|---|---|
| Microsoft Copilot | Drafting and summarising inside Office 365 | Around £19 per user per month on top of your Microsoft plan | Value depends on already using Microsoft 365 |
| ChatGPT (OpenAI) | General drafting, brainstorming, coding help | Free tier; paid plan around £20 per user per month | Keep confidential data out unless on a business plan |
| Google Gemini | Businesses on Google Workspace | Add-on pricing per user, varies by Workspace plan | Best value only if you are already Google-based |
| Claude (Anthropic) | Long documents, careful writing | Free tier; paid plan around £18 per user per month | Fewer built-in app integrations than rivals |
| Zapier / Make | Connecting apps so data flows automatically | Free tiers; paid plans from roughly £15 to £25 per month | Costs rise with volume; plan your “tasks” budget |
| Synthesia | Training and explainer videos from a script | Paid plans from around £18 per month | Overkill unless you make video regularly |
Prices move often, so always check the vendor’s own page before you commit. Treat the figures above as a shortlist starting point, not a quote.
A realistic worked example
Picture a six-person plumbing and heating firm in Leeds. The owner spends most Friday afternoons writing quotes and chasing unpaid invoices. Here is how a front-runner approach would play out.
First, quoting. The owner sets up a Copilot or ChatGPT prompt that turns rough job notes into a tidy, branded quote. What was 20 minutes of typing per quote becomes five minutes of checking. Second, invoice chasing. The firm already uses accounting software, so instead of writing polite reminders by hand, they lean on the automated payment reminders built into tools like Xero, QuickBooks or the UK-built FreeAgent. Third, they use Make to move a completed job in their scheduling app into a “raise invoice” task automatically, so nothing slips.
The result is not science fiction. It is a couple of hours a week back, plus fewer overdue invoices. If cash flow is your pressure point, pair this with our advice on improving cash flow in a small business, because time saved on admin only helps if it turns into faster payment.
Where SMEs go wrong
The slower adopters in this kind of research usually trip on the same things. They buy a licence and never build it into anyone’s daily routine, so it quietly lapses. They try to automate a messy process instead of tidying the process first. Or they skip the human check, get burned by one confident-but-wrong answer, and abandon the whole idea.
There is also a governance side. Putting customer data or confidential figures into a free consumer AI tool can breach your obligations under UK GDPR, so it pays to read the guidance from the Information Commissioner’s Office and to use business-tier plans that keep your data out of training. Our overview of AI tools for small businesses covers how to weigh these trade-offs.
Frequently asked questions
Do I need technical skills to become an AI front-runner?
No. The drafting tools work in plain English, and platforms like Zapier and Make use drag-and-drop steps rather than code. The skill that matters is process thinking: knowing which job to point the tool at and how to check its output.
How much should a small business budget to start?
You can begin for very little. A single paid assistant seat is roughly £18 to £20 a month, and most automation tools have a free tier for low volumes. A sensible starting budget is under £50 a month across one assistant and one automation tool, then scale up only when you have measured a real time saving.
Is AI safe to use with customer data under UK rules?
It can be, if you use business or enterprise plans that state your data is not used to train models, and if you avoid pasting sensitive information into free consumer tools. Follow the ICO’s AI guidance and keep a simple record of which tools handle what data.
Which single tool should I try first?
If you already pay for Microsoft 365, start with Copilot because it works where you already do. If you live in Google Workspace, start with Gemini. If you use neither heavily, a paid ChatGPT or Claude plan is the flexible choice. Do not buy all of them at once.
How do I know if it is actually saving time?
Time one task before and after. Write down “this took 90 minutes, now it takes 30” for the first few weeks. That crude measurement is exactly what the front-runners do, and it tells you whether to expand or stop.
What to do next
- Pick one task this week. Choose the most frequent, boring, low-judgement job you do and write down how long it takes now.
- Match one tool to it. Use the comparison table above to shortlist a single assistant or automation tool, and start on a free or entry plan.
- Keep a human check and measure the result. Review the AI’s output before it goes out, and record the time saved after two weeks.
- Expand only what works. Once one task is genuinely faster, repeat the exercise on the next task rather than buying more tools you never embed.
