AI for SMEs – What I Learnt and What You Should Know

If you run a small business (like us at Innovate Ltd) with an even smaller IT team, you’ve probably been told at least once this month that AI will transform your operations. Maybe your organisation has dabbled in it, you don’t quite know where to start, your staff are getting frustrated because they’re not using effectively, or you simply struggle to prove ROI. I managed to spend £3k in one week alone on Claude because I forgot to watch a meter, so I have first-hand experience (and advice on what not to do…)

My newsletter (originally published on LinkedIn but also featured on our website here) will answer the questions: “What can I use AI for?”, “How can I introduce AI safely and effectively as part of an organisational transformation strategy?”, and “How can I prove the ROI of AI” – from one director to another, with my insight as someone who’s worked in IT for over two decades but in a language we both understand.

To keep this first edition brief, I’m outlining the challenges as I see them at the moment, and I’ll dive deeper into each of them as this newsletter series progresses. Me and my team will also be holding on-theme workshops and webinars for some thought-provoking hands-on insight from successful (or even failed) AI implementations you can learn from, so look out for those announcements.

The Top-Level AI Challenges How I See Them

Finding the right use case and the right model.

In my view, the most common mistake is starting with the technology for the sake of it (“we need AI”) rather than the problem (“our quotes take three days”), so pick one repetitive, measurable pain point first and only then choose the tool – a general-purpose assistant such as Copilot or ChatGPT for office tasks, or a specialist supplier who understands your industry for anything more complex.

Using AI ethically – supporting people, not replacing them.

In a small business, your team is your competitive advantage, so I firmly believe AI should take on the tedious tasks rather than full-on replace people – or making people’s lives more difficult.

Taking your staff with you – change management is key.

If staff suspect the introduction of AI is a cost-cutting exercise in disguise, they will resist it. Not only that: a lack of skills is consistently named as one of the biggest barriers to adoption in UK businesses. In my experience, what works is explaining the “why” early and honestly, involving the people who do the work in choosing and testing tools, appointing one or two enthusiastic “AI champions”, investing in practical training rather than just software licences, and sharing quick wins – giving credit to the team. Change management is a two-way street.

Proving return on investment.

This has admittedly been one of the most challenging aspects so far. Benefits such as time saved are real but hard to quantify while costs are easy to see. We’ll be looking into more quantifiable results rather than vanity metrics.

Protecting your data.

There’s a fine line between AI making use of your data to your benefit and leaking corporate secrets. Anything you type into an AI tool may be stored – and some free consumer tools may use it to train future models (I’ll talk about so-called Shadow AI – non-sanctioned use of personal AI tools to process company data – in an upcoming newsletter). Under UK GDPR, you remain responsible for any personal data about customers or employees that you share, and commercially sensitive information such as pricing, contracts and product designs deserves the same care. Think of AI as Industry Espionage 2.0.

Posing new dangers.

AI does not only help your business – it supports criminals in their work, too. Attackers can hide instructions in emails, PDFs or web pages that trick AI assistants into leaking data or taking actions. That’s not even to speak of the danger of digital clones tricking staff into divulging secrets. We talked about those dangers in our recent webinar if you’re interested.

As you can see, AI is a complex beast, but I’m looking forward to exploring these topics in-depth and get the conversation going.