Four in five UK firms can't prove their AI is paying off. Write the measure down before you spend

The good news, published the same week, is that the gap is closable, and the fix isn't a better tool. It's three plain habits.

How firms are falling into the gap

The Air IT findings read like a list of avoidable mistakes:

• 41% bought or introduced an AI tool before agreeing what problem it should solve.
• Among those with active AI projects, 74% said not all of them had both an agreed objective and a measure of success.
• 47% said AI had made at least one task slower once checking, correction, set-up and management were counted.
• Only 13% consistently turn time saved by AI into measurably higher output.
• Of organisations that formally track pilots, 62% abandoned at least one in the past year, most often because of integration difficulties (34%) and staff not adopting the tools (29%).

None of that says AI doesn't work. It says AI was bought like a gadget rather than introduced like a change to how the business runs.

The firms that can show a return do three simple things

In an essay for the British Chambers of Commerce's new collection How the UK can become an AI Adoption Superpower (5 November), Richard Breeden of aibl draws on a survey of 755 UK mid-market leaders. The share of firms able to show a measurable return climbs from 22% with no AI governance to 85% where governance is mature, using much the same tools.

"Governance" here isn't a committee. It's three habits:

1. Someone owns the outcome.
2. The measure of success is written down before the money is spent.
3. The numbers hold up when the finance director checks them.

Enthusiasm alone doesn't close the gap: only 39% of "all-in" firms without governance can show a return, against 54% of well-governed firms that aren't especially keen. Pair governance with capable people and 81% show a return. Leaders matter too: 71% of firms whose leaders use AI in their own work can show a return, against 31% where leaders barely touch it.

Those figures come from larger mid-market firms (£50m to £500m revenue), but the habits scale down perfectly. A ten-person business can name an owner and write a target on one page.

The quiet blocker: systems that don't talk to each other

Integration was the top reason pilots were abandoned, and it isn't just an AI problem. Research for Finexer, reported on 6 November, found 27% of UK businesses with 10 to 500 staff wait more than nine weeks for banking or payment integrations to go live. Only 10% have highly automated banking and payment workflows, and 39% had staff tracing, comparing or re-entering payment information across systems in the past year.

That matters because AI can only act on data it can reach. If the order, the invoice and the bank line live in three systems joined by copy and paste, an AI assistant just adds a fourth place to check. With Making Tax Digital rolling on and mandatory electronic VAT invoicing planned from April 2029, connected financial systems are becoming a necessity rather than a nice-to-have.

Good advice helps, too

A November policy briefing from the Enterprise Research Centre at Warwick Business School, using the government's Longitudinal Small Business Survey, adds a longer-term view. Only around 6.8% of UK SMEs adopted advanced technologies (AI, robotics, automation and similar) between 2018 and 2021. SMEs that used external information and advice were more likely to adopt them, particularly when that advice came from professional advisers or business and education networks and focused on efficiency or innovation. External advice also remained positively associated with labour productivity across all the researchers' models.

A one-page plan to close your own decision gap

1. Start with the problem, not the product. Write one sentence: "We want to reduce X from A to B by date C." If you can't, don't buy yet.
2. Name one owner. Not IT, not "the team". One person who reports the result.
3. Baseline before you switch anything on. Time per task, error rate, days to cash, response time. Two weeks of honest numbers is enough.
4. Count the hidden time. Include checking, correcting and managing the AI, which is exactly what made tasks slower for 47% of Air IT's respondents.
5. Fix the connections first. Decide which system holds the truth for customers, orders, invoices and payments, and connect the others to it before layering AI on top.
6. Plan adoption like a launch. Train the people who'll use it, show leaders using it, and agree what happens to the time saved.
7. Review at six to eight weeks, then decide. Scale it, fix it or stop it, and record why. Stopping a pilot with evidence is a success; drifting is the failure.

The optimistic bit

UK businesses aren't short of ambition or tools. The research this week shows the difference between AI that pays and AI that quietly costs money is mostly organisational: an owner, a written target and systems that share data. Those are cheap to put in place, and once they exist, every future digital project gets easier to judge.

Why this matters for Clarity Growth

Clarity Growth helps UK SMEs and scale-ups with exactly the parts of digital transformation that close the decision gap: defining the problem and success measure before tools are chosen, mapping the process, deciding the system of record, and building connected systems so websites, ecommerce, CRM, finance and operations share clean data. Then we add automation and AI where the baseline says it will help, with logging and approvals, and a review point that tells you whether to scale it. It's practical, outside advice of the kind the ERC research links to better adoption, delivered as small, measurable projects rather than another subscription.

Sources

1. Four in five UK businesses struggle to prove AI investment is paying off, Air IT Group research finds — Thames Valley Chamber of Commerce (5 November 2026; 504 UK decision makers: 83%, 41%, 74%, 47%, 13%, 62%, 34%, 29%)
2. Firms struggling to prove AI investment is paying off, research finds — Greater Birmingham Chambers of Commerce (9 November 2026; coverage of the same Air IT Group research)
3. Governance and capability decide which UK firms make AI pay — aibl (5 November 2026; Richard Breeden's essay for the British Chambers of Commerce and BDO collection; aibl State of UK AI Adoption Survey 2026, 755 mid-market leaders: 22% vs 85% return by governance, 81%/40%/15% by governance and capability, 71% vs 31% by leader use)
4. How the UK can become an AI Adoption Superpower — British Chambers of Commerce (PDF) (November 2026; 29-essay collection with BDO)
5. More than a quarter of firms face nine-week bank delays — CFOtech UK (6 November 2026; Finexer research, 418 leaders at UK firms with 10 to 500 staff: 27% wait 9+ weeks, 10% highly automated, 39% re-entering payment data; e-invoicing from April 2029)
6. Advanced technology adoption in UK SMEs: policy briefing — Enterprise Research Centre (PDF) (November 2026; LSBS analysis: ~6.8% AET adoption 2018–2021; external advice associated with adoption and labour productivity)