Indeed’s mid-year UK labour market report, published in early August 2026, found that 9.4% of UK job postings now explicitly mention AI or AI-related tools — a record high, and a figure that’s climbed sharply in just a few years. In software development and data roles, almost half of vacancies now reference AI. The twist is that this is happening inside a shrinking market: overall UK job postings are down 11% since the start of the year and sit roughly 32% below pre-pandemic levels. Employers are posting fewer roles overall, but a growing share of the roles they do post now expect some form of AI capability.

For UK SME owners hiring even occasionally, that’s worth pausing on — not because you need to chase a trend, but because “AI skills” has become a phrase people put in job ads without necessarily knowing what they mean by it, on either side of the interview table.

The other detail in Indeed’s report worth noting: jobseeker searches for AI-related roles have risen sevenfold since ChatGPT launched, and that interest is no longer confined to software and data roles. HR, marketing, finance and management job ads are all starting to reference AI tools, which means the “AI skills” question is now landing on SME owners who’ve never had to think about it before — not just tech firms with a dedicated recruiter who already knows the difference between a data scientist and someone who’s handy with a chatbot.

What “AI skills” actually means in practice

There’s a wide gap between the two things that phrase can mean. At one end is deep technical capability — building, fine-tuning or deploying AI models, relevant mainly to specialist technical hires. At the other is practical fluency — knowing how to use tools like ChatGPT, Copilot or Claude effectively for research, drafting, data analysis or customer service, which is now genuinely useful across marketing, HR, finance and operations roles, not just engineering.

Most SME roles need the second thing, not the first. But job ads that just say “AI skills desirable” without specifying which kind tend to either scare off good candidates who assume they need a technical background they don’t have, or attract candidates who overstate familiarity with tools they’ve barely used. Neither helps you hire well.

Writing job ads and interviews that actually work

Name the tool, not the buzzword. “Comfortable using AI tools like Copilot or ChatGPT to draft and research” tells a candidate far more than “AI-literate,” and it’s easier to test for honestly at interview.

Ask for a specific example, not a self-rating. “Tell me about a time you used an AI tool to solve a real work problem” filters out inflated CV claims far better than asking someone to rate their own AI skills.

Separate “uses AI tools” from “understands AI limitations.” For roles touching customer data, finance or compliance, the more valuable skill is often knowing when not to trust an AI output unchecked, rather than raw enthusiasm for using one.

Don’t assume existing staff need external hires to close the gap. A short, structured internal AI upskilling push is often cheaper and faster than hiring for it — and it’s exactly the kind of practical, business-first AI adoption work BuildApps specialises in for UK SMEs that want capability without the buzzwords.

Be honest about what the role doesn’t need. Not every hire benefits from AI fluency, and padding a job ad with it anyway just adds noise for candidates and makes genuinely AI-relevant postings harder to spot. If the role is mostly manual, physical or relationship-driven work, leave it out rather than following the trend for its own sake.

The takeaway

A record share of job ads now mention AI, but that doesn’t mean every SME role needs a specialist. Get precise about which kind of AI capability the role actually needs, write the ad and interview questions around that specific thing, and you’ll hire better — and avoid either scaring off good candidates or overpaying for skills you didn’t actually need.