“I’ve got a million questions about AI. I just haven’t worked out what they are yet.”
That’s more or less what one business owner admitted at our AI Lunch & Learn back on the 2nd of July, and going by the nods around the room, she wasn’t alone. We’d got a room full of owners and senior leaders from businesses across Essex together, the sort of firms we work with every day, for an honest session on what AI actually means for a company their size.
What stuck with us afterwards was how practical the questions were. They wanted to know where to start, whether it was safe, and what it was going to cost them. So here’s what business owners around Basildon and the wider Essex area actually wanted to know about AI, and the straight answers we gave.
“Am I already using AI without realising it?”
Here’s a question that caught a few people out. When we asked the room who was using AI, a handful of hands went up. When we asked who had staff using AI, a few more went up gradually. And when we asked whether those tools were free or paid, the honest answer for most was “the free ones, and nobody really signed them off.”
That’s shadow AI, which we addressed in an article earlier this year, and it’s already in your business whether you’ve planned for it or not. It matters because free tools aren’t really free. You’re just paying with your data instead, and personal data fed into a public AI tool is still personal data as far as the ICO is concerned. “We didn’t know it kept the information” isn’t much of a defence after the fact when it’s already too late.
Thankfully, the fix is a single page. A one-page AI policy that sets out which tools are approved, what can and can’t go into them, who to ask, and who owns it is enough to get a grip on the risk. It turns out most UK firms are in exactly this spot: 54% now use AI in some form, up from 35% a year earlier, but only around one in ten use it in any structured way. The rest are experimenting, often without realising it.
“Where do I actually start, and what does it cost?”
Once people accept AI is already in the building, the next question is always the same: where do I start, and what’s it going to cost me?
The honest answer surprised a few people, because the best place to start is deliberately boring. The wins that give staff real time back usually come in the form of drafting emails, proposals and client updates, summarising long documents and meeting notes, tidying and checking spreadsheets, and pulling first drafts of marketing together. They’re just ordinary jobs, done faster. As we put it on the day, one hour a week saved across a team of twenty adds up to nearly 960 hours a year. That’s most of a working year handed back.
On cost, the reframe that landed hardest was this: you may already be paying for it. If you’re on Microsoft 365, tools like Copilot keep your data inside your own tenant, under agreements you’ve already signed, which makes them a far safer starting point than yet another free tool nobody vetted. The trick is picking one use case, proving it works in a few weeks, then adding the next. Start small, prove it, and build from there.
“How do I make sure I don’t waste money on this?”
This was the wariest question in the room, and rightly so. Plenty of businesses have spent money on AI and got nothing in return, so the session attendees wanted to know how to avoid being one of them.
The reasons AI projects fail are fairly predictable: someone buys a tool before anyone has agreed what problem it’s solving, then the tools pile up and you end up with three that do roughly the same thing and one that nobody opens. There’s the trust gap, where you spend so long checking the output that it eats the time it was meant to save. And then the silent stall, where nobody owns it, nobody reviews it, and nobody quite decides to stop. Gartner expects 60% of AI projects to be abandoned by the end of 2026, largely because the underlying data was never ready for them.
The way to avoid all of that is a bit of structure, not more spend. Pick the use case. Give one person ownership of it. Choose a tool that fits your data and your appetite for risk, rather than whatever’s trending. Then actually review it, at two weeks, four weeks and sixty days, and decide whether to keep going, change tack or drop it. It doesn’t need to be a grand transformation programme. It needs an owner and a date in the diary.
Why this conversation is different in Essex
Most of what’s written about AI isn’t written for a business like yours. The big national coverage assumes a company with a data team, a training budget and someone whose actual job is “AI”. That’s not the room we were sitting in. The room was full of Essex firms with fifteen or twenty staff, the architecture practices, accountants and professional services businesses that make up the backbone of the local economy, where the person who’d end up owning AI is already doing three other jobs.
For a business that size, the barrier is usually skills rather than money. That’s precisely why a session where you can ask a real question and get a straight answer beats another large-scale think-piece. It’s also why the setting mattered. There’s a reason the most-quoted UK research on business AI adoption was pulled together with the University of Essex: this conversation has always been closer to home than you may have realised.
And that’s really the point of doing it in person, locally. AI for a Basildon business is a far more useful conversation when the people answering your questions actually understand a business your size and are based just down the road.
Want the full picture?
We covered a lot more on the day than we could fit here, from getting Microsoft 365 Copilot working safely to the agentic tools that are starting to do whole jobs rather than just answer questions to a straight look at how AI is changing the cyber security picture for smaller firms.
If you’d find that useful, you can download the full presentation from our AI Lunch & Learn and get all the detail we went through on the day.
And if you’d rather talk it through for your own business, that’s what we’re here for. Have a proper conversation with someone who understands a business your size and is based just down the road.
FAQs
Is it safe for a small business to use AI?
It depends on the tool. Free, public tools can learn from whatever you type in, so keep anything sensitive out. Tools like Microsoft 365 Copilot are safer because your data stays in your own environment. Best first move: a short AI policy setting out which tools are approved and what can go into them.
How should a small business actually get started with AI?
Start small. Pick one repetitive task that eats time (something like drafting emails, summarising documents, or tidying spreadsheets) and give one person ownership of trying AI on it. Prove it works over a few weeks, then move to the next.
Do I need a big budget for AI?
Not usually. Many of the best starting points are tools you may already pay for, like Microsoft 365 Copilot.
What’s the biggest mistake businesses make with AI?
Buying a tool before agreeing what problem it’s meant to solve. It’s why so many AI projects stall and why Gartner expects organisations to abandon 60% of AI projects through 2026. Decide the use case first, give someone ownership, then review whether it’s working.
Can Outbound help my business in Basildon get started with AI?
Yes. Outbound is based locally and works with businesses across Basildon and the wider Essex area. We’ll help you find where AI genuinely fits, keep it safe, and start without wasting money.
