Then the hangover. Only 14% said they have comprehensive, real-time visibility of AI costs. Almost a third (32%) reported significant gaps in their ability to assess return on investment. Productivity is being claimed. The invoice trail is still fuzzy.
For UK SMEs and scale-ups, that gap matters more than the headline. Large-firm CEOs can bury opaque AI spend inside bigger budgets. Growing businesses cannot. If you cannot name what AI costs and what it changed this month, you are not "adopting AI" — you are collecting subscriptions.
What EY actually found
The UK pulse sits inside EY's global CEO Outlook work (1,200 CEOs surveyed August–September 2026; 100 UK respondents). Most of those UK companies are large: 72% reported annual revenue over US$1 billion. Treat the productivity claim as a direction of travel for serious operators, not a mirror of a 12-person Northampton team.
Still, the blockers translate cleanly:
| Signal (UK CEOs) | Figure | Why SMEs should care |
|---|---|---|
| AI among top-3 productivity drivers | 50% | Vendors will keep selling "productivity"; your job is proof |
| Comprehensive real-time AI cost visibility | 14% | If big firms struggle, your credit dashboards need owners |
| Significant ROI assessment gaps | 32% | No metric = no kill switch for weak tools |
| Biggest obstacle to converting gains into value | Legacy tech / operating models 34% | New AI on broken process is theatre |
| Skills shortages > capital as growth barrier (next 3 years) | 74% | Budget seats without training and you stall |
| Not developing workforce skills fast enough for AI | 47% | Time for one owner beats another pilot licence |
EY also reports how leaders feel the productivity: greater output without more headcount or capital (34%), faster delivery (19%), more time for higher-value work (15%). More than two-fifths (44%) say they are reinvesting those gains into growth, innovation or transformation. That is the healthy pattern — if you can measure it.
Silvia Rindone of EY-Parthenon UK & Ireland framed the shift bluntly: the conversation has moved from adoption to impact. Matthew Ringelheim, EY UK & Ireland AI & Data Leader, added the same caution on deals: AI can help evaluate transactions, but it is no substitute for fundamentals. For an SME, swap "transaction" for "tool purchase" and the advice holds.
What ONS says about everyone else
EY speaks for large-company CEOs. The Office for National Statistics speaks for the wider private-sector picture.
In its 8 October 2026 Business Insights bulletin (Wave 165, late September field period), ONS reported that 30% of all businesses used at least one AI technology in September 2026 — up from 9% when the question launched in September 2023, and broadly stable versus June 2026. Size still dominates:
- 51% of businesses with 250+ employees use AI
- 29% of businesses with fewer than 10 employees use AI
- For firms with 10+ employees, AI use sits at 38% (up 3 points from June)
Among 10+ employee AI users, the most common technologies remain text generation with large language models (21%) and visual content creation (19%). The roles most affected are still administrative and clerical (38%), then creative or design (24%). Only 6% of those AI-using firms reported reduced headcount as a result; fewer than 1% reported headcount up. AI is rearranging work more than it is resizing the payroll — which matches EY's finding that 71% of UK CEOs expect AI to change roles and skills more than workforce size.
Industry skew is real: 60% of information and communication businesses report AI use — the highest of any surveyed sector. If you sell services online, you are competing in a market where a large minority already treat GenAI as normal ops kit.
Productivity without a ledger is a story, not a strategy
The EY–ONS pairing gives UK growing businesses a useful tension:
- Leaders believe AI is already a top productivity lever (EY).
- Most still cannot see cost or ROI clearly (EY).
- Adoption outside the biggest firms is real but uneven, and concentrated in text/image tools hitting admin work (ONS).
That is why "we use ChatGPT" is not a transformation update. A 14% cost-visibility score among CEOs of large UK companies is a warning light for every SME buying Copilot seats, image tools, agent plugins or vertical AI add-ons before peak season.
Legacy operating models (34% in EY) are the other half of the story. If order status still lives in three spreadsheets, an LLM that drafts nicer emails will not move contribution margin. Clarity Growth clients see this constantly: the model is fine; the workflow, data and measurement are not.
A practical checklist for UK SMEs this month
Use EY's 12 October numbers and ONS's size gap as a forcing function — not another reason to panic-buy.
- Open one AI cost ledger. Seats, credits, API usage, vendor AI add-ons. If finance cannot total it in 15 minutes, you are in the 86% without clear visibility.
- Pair every tool with one outcome metric. Time saved on a named workflow, ticket deflection, content cycle time, quote turnaround — pick something you can read in 30 days.
- Start where ONS says work is already shifting. Admin and clerical tasks are the modal impact. Automate handoffs, FAQs, first drafts and status updates before you fund speculative "agent" projects.
- Fix the operating model before you stack models. EY's top value blocker is legacy tech and ways of working. Map the process the AI is meant to touch; remove double entry; then connect the tool.
- Budget skills as a line item. 74% of UK CEOs expect skills shortages to beat capital as a growth constraint. Protect hours for one owner to review outputs, kill bad automations and train the team.
- Prefer task-packaged AI with export paths. Narrow jobs (reception, cataloguing, support drafts) beat blank-canvas GenAI when you lack an AI lead — but know what you can move if the vendor changes pricing.
- Reinvest only measured gains. EY's 44% reinvestment story is the goal: free capacity should fund growth work, not more unused licences.
Why this matters for Clarity Growth
Clarity Growth helps UK SMEs and scale-ups turn AI curiosity into connected, measurable operations — websites, ecommerce, content systems and the tools that feed them. EY's 12 October survey confirms the market mood: leaders already feel AI in productivity. The same survey shows why most firms will waste money without cost visibility, process redesign and skills. ONS shows smaller firms still lag on adoption, and that GenAI is landing hardest on admin work. We help clients choose where a packaged AI tool fits, design human review and escalation, stitch it into real CRM, commerce and support flows, and put a simple ROI ledger around spend so "productivity" becomes a number the board can trust — not a slide.
Sources
- AI drives UK productivity gains, but barriers to growth remain — EY UK (12 October 2026)
- AI drives productivity gains for UK businesses but barriers to growth remain — Business Money (12 October 2026)
- Business insights and impact on the UK economy: 8 October 2026 — Office for National Statistics (8 October 2026)