Where AI Actually Fits in Your Business: A Framework for UK SMEs

OpenAI just raised $110 billion. Biggest private funding round in history. Amazon, Nvidia, and SoftBank put up the money. The company is now worth $730 billion.

And yet a recent YouGov survey found that 30% of UK SMEs still can’t identify a single place AI would be useful to them.

I don’t think that’s true. I think every business already knows where AI fits. They just don’t recognise it yet.

Here’s what I mean.

Your business is already full of systems

Every business has systems. Not software. Human systems. Processes that run on people doing the same thing the same way every week.

Someone copies data from one place to another. Someone reformats a report. Someone chases a client for an update. Someone screens enquiries and sorts them into categories. Someone takes information from an email and types it into a spreadsheet.

Nobody calls these “systems.” They call them “just part of the job.” But that’s exactly what they are. Manual systems running on human effort instead of automation.

The question isn’t “where could AI help my business?” That’s too abstract. Most people hear that and freeze because it sounds like a massive decision. It sounds like they need to understand the technology, choose the right platform, hire someone technical, and hope it works.

That’s where AI fits. Every single time.

Not replacing people. Replacing the robotic work that people are doing between systems that don’t talk to each other.

Invisible systems hiding in plain sight

These bridges between systems look different depending on the sector, but the pattern is always the same. A person is doing something repetitive, structured, and predictable because two tools or processes don’t connect.

Recruitment agencies

Someone reads a CV, pulls out the relevant details, and types them into the CRM. Someone else scans a job spec, matches it against a list of candidates, and sends out template emails. Someone reformats interview notes into a client-facing summary. According to StandOut CV, recruiters who automate these gaps fill 64% more vacancies than those who don’t. Not because they hired more people. Because they stopped using people as glue.

Accounting firms

Someone receives bank statements by email, downloads them, opens a spreadsheet, and manually categorises transactions before entering them into the accounting software. Someone else pulls data from three different reports to build a quarterly summary for a client. Someone chases clients for missing documents every single month, sending the same email with slightly different dates.

Law firms

Someone reads through a contract and highlights the key clauses to summarise for a partner. Someone compiles case notes from emails, calls, and documents into a single brief. Someone checks every new enquiry against conflict-of-interest criteria by searching through old client files.

Property agencies

Someone takes photos and descriptions from a viewing and writes up the listing in three different formats for three different portals. Someone logs tenant maintenance requests from emails and phone calls into a tracking system. Someone cross-references tenancy dates against renewal deadlines and sends reminders manually.

HR consultancies

Someone drafts a policy document by pulling from templates and adjusting the details for each client. Someone tracks which clients are due for annual reviews and sends reminder emails. Someone takes notes during a disciplinary hearing and turns them into a formal report.

Trades businesses

Someone receives an enquiry by phone, writes it down, and later enters it into a quoting system. Someone checks the diary, matches availability to job requirements, and sends confirmation texts. Someone compiles timesheets from paper records into a payroll spreadsheet.

The pattern is always the same. A person sitting between two things, manually moving information from one format or location to another. The work isn’t creative. It isn’t strategic. It’s connective tissue. And it’s the exact kind of work that AI handles well.

How to spot the invisible systems in your business

Most business owners struggle to identify these processes because they’ve been normalised. They’ve been “just part of the job” for so long that nobody thinks to question them.

Here are five questions that surface them almost immediately.

• Where does someone in your business copy information from one place to another? This is the most common invisible system. Data lands in one format (an email, a form, a document) and a person manually transfers it into another format (a spreadsheet, a CRM, accounting software). Every instance of this is a process that could run itself.
• What task does your team do every week that follows the same steps in the same order? If the steps are predictable, the task is automatable. Not every repetitive task needs AI, but every repetitive task deserves the question: why is a person still doing this?
• Where does your team chase people for information they should already have? Chasing clients for documents. Chasing suppliers for updates. Chasing internal teams for status reports. If the chasing follows a pattern (send reminder, wait, follow up, escalate), that’s a system pretending to be a conversation.
• What would break if one specific person was off sick for two weeks? This usually reveals processes that live entirely in someone’s head. They know which clients need chasing, which reports are due, which leads haven’t been followed up. That knowledge isn’t a system. It’s a risk. And it’s exactly the kind of thing that should be captured, structured, and partially automated.
• Where does your team spend time reformatting or repackaging the same information for different audiences? Internal reports turned into client-facing summaries. Job specs adapted for different job boards. Meeting notes restructured into action lists. Every time information is reshaped rather than created, that’s a bridge a person is acting as.

If you ask yourself those five questions honestly, you’ll find at least three or four processes that are running on human effort when they don’t need to be.

The compounding effect

Here’s what I’ve seen with every business I’ve worked with. Fixing one process changes how you see the rest.

Before the first automation, business owners tend to think in terms of tools. “We use Xero for accounting. We use HubSpot for CRM. We use Google Workspace for email.” Each tool is a separate thing. The gaps between them are invisible because they’ve always been there.

Once you connect one gap, you start seeing the others. You fix the way enquiries come in, and suddenly you notice how much time is wasted on the next step. You automate the reporting, and you realise the data feeding into those reports is being entered manually. You streamline client onboarding, and it becomes obvious that the follow-up process is held together with sticky notes and memory.

That’s how it compounds. Not because you bought more technology. Because you trained your eye to see the systems that were always there.

The businesses pulling ahead right now aren’t the ones with the biggest AI budget. They’re the ones that spotted one broken process, fixed it, and kept going. Each fix is small on its own. But the cumulative effect of removing three, four, five manual bridges from your operations is significant. Your team gets faster. Your clients get a better experience. Your capacity grows without your headcount growing.

And it starts with just one.

The technology exists. The bottleneck is seeing it.

The $110 billion in funding proves the technology isn’t going anywhere. The tools are mature enough to deploy in a 10-person firm today. The cost isn’t prohibitive. The bottleneck is something much simpler.

It’s seeing the invisible systems already running inside your business.

Once you see them, you can’t unsee them. Once you fix one, you start seeing the next. And that’s the real competitive advantage. Not the AI itself. The ability to look at your own operations clearly and spot where human effort is being used as glue between things that should already be connected.