Bank of England Finds AI Drives a Fifth of UK Growth for a 0.5% Productivity Gain

TL;DR

The Bank of England's November Monetary Policy Report, published November 5 alongside an 8-1 vote to raise Bank Rate to 4%, puts an official number on AI's slice of the UK economy: novel AI-related activity accounts for "just under a fifth of GDP growth since the start of 2026". Almost all of that is upstream, from firms building and installing AI systems. Surveyed firms put their own productivity gain at about 0.5% over three years, report close to zero effect on jobs, and expect 0.9 points a year from here. The Bank kept its assumptions where they were and filed the upside as a risk.

How the Bank counts AI

UK national statistics have no line item for AI, so Bank staff built an "AI activity proxy" from three standard industrial classifications: SIC 58, publishing activities, which in the UK taxonomy includes developing generative AI agents and AI application software; SIC 62, computer programming and consultancy; and SIC 63, information services, home to data processing and data centres. Strip out the longer-run drift of tech activity, and what remains produces the headline: "novel AI-related activity accounts for just under a fifth of GDP growth since the start of 2026."

The Bank admits the seams. "Not all the output of these sectors will capture AI activity," the report says, and publishing also contains books and newspapers. In the UK's filing system a generative agent and a romance novel share a drawer. The report leans on the Office for National Statistics, which notes that many of the businesses reporting the strongest turnover growth in these industries are involved in AI and cloud computing.

The productivity number firms report

The survey numbers come from the Decision Maker Panel, the Bank's poll of company executives. Firms estimate AI has increased their productivity "by around 0.5% over the past three years, in line with similar impacts reported in other countries". That is 0.1 to 0.2 percentage points of annual productivity growth since early 2024.

The report lists three candidate explanations for the recent productivity pickup: cyclical adjustment as firms unwind post-pandemic labour hoarding, cuts to lower-paid roles after rises in the National Living Wage and employer National Insurance, and AI, which would imply a structural and persistent shift. The Bank's verdict is that "much of the recent strength in productivity growth appears to reflect cyclical adjustments". Productivity has improved in sectors with little AI exposure, dispersion across industries is no higher than before Covid, and even information and communications, the most AI-heavy sector, has productivity growth that "has not risen above pre-Covid averages".

Fewer hires, same firings

For anyone staffing a team, the employment finding is the sharpest part. In aggregate, DMP firms "report close to zero impact of AI on employment so far". The mechanism comes from the Bank's regional Agents: AI adoption "is reducing the likelihood of firms hiring or replacing workers, but not materially increasing the likelihood of firing workers". Some contacts "have reached their lower bound for headcount". Nobody is being shown the door by a model; the vacancies just go unfilled. Unemployment stood at 4.7% in the three months to August.

Demand now, supply later

A central bank cares because of inflation arithmetic. Spending on data centres and AI services is demand, and demand that outruns supply pushes on prices. Productivity gains would be supply, letting the economy grow faster without inflation. On the evidence so far the Bank puts the UK in the first camp: "the roll out appears to primarily reflect a boost to demand within the economy." The contribution to growth "appears to reflect activity in the upstream development and implementation of AI technologies rather than broader gains from their adoption elsewhere in the economy," a pattern the report calls "consistent with the impact on productivity having been relatively small so far."

The activity surprise shows in the GDP numbers. Growth was 0.5% in the second quarter and is expected at 0.7% in the third, against a July projection of 0.3% and 0.1%. Business-to-business services, where much AI work lives, contributed 0.6 points of the 0.7% growth in the three months to August. Business investment rose 3% in the first quarter and 1.8% in the second, the fastest two quarters since 2022, and the Agents point to AI-related spending as a driver. Counting a build-out as transformation is like counting a construction site as housing: the cranes are in the accounts, the finished flats are not.

The memory-chip tax

AI also reaches Britain through prices. "Demand for AI-related products such as memory chips has placed upward pressure on export prices, particularly in China and advanced Asian economies," the report states, building on analysis in its July edition. For Chinese exports, AI-related goods saw the largest price increases, according to the report's risk section. UK import prices excluding fuel are projected to rise 1.9% in the year to the fourth quarter, against 0.8% expected in July, though that revision carries energy costs as well. The same pressure already sits in your hardware invoice; the Bank has now booked it nationally.

What firms expect, and what the Bank assumes

The gap between experience and expectation is the widest number in the report. DMP firms expect AI to add "about 0.9 percentage points per year on average over the next three years, split roughly equally between slower employment growth and faster output growth". For scale, 0.9 points is close to the MPC's entire 1% assumption for trend productivity growth.

The Bank left the assumption at 1% and called AI an increasing upside risk instead, with a review promised once the ONS publishes re-based productivity figures on November 17 and industry-level data in February. Outside the UK the picture has the same shape. A Gartner survey in October found 3% of organizations had met their AI productivity or cost targets, 38% were already over budget, and 85% planned to spend more anyway. The IMF warned last month that the spending spree could drag on the wider economy if promised profits fail to appear.

What to steal

• If you sell to AI builders, you are inside the fifth. Consultancies, data services, agent shops: the national accounts say that is where UK growth concentrated this year.
• If you sell AI to everyone else, the measured payoff is 0.5% over three years. Sell the specific task you remove; leave macro promises out of the deck.
• Budget hardware inflation. The same demand lifting Asian export prices lifts your server quote, and the Bank now projects import prices rising faster than it thought in July.
• Plan hiring as the adjustment margin. The Bank's contacts fill fewer roles and replace fewer leavers, and firms expect half the coming productivity gain to arrive as slower hiring.
• Watch November 17 and February. When the ONS re-bases the productivity data, every AI ROI argument gets either new ammunition or a new problem.

Caveats

• The "fifth of growth" figure rests on the proxy: three industry codes full of non-AI output, adjusted by a trend assumption the report does not quantify. It is a Bank staff estimate, not an ONS statistic.
• The 0.5% and the 0.9-point figures are self-reported by surveyed firms; the report gives no sample size or question wording, and expectations are not measurements.
• This is UK data only, and the ONS is mid-way through changing how it measures labour productivity, so the baseline itself is moving.
• The Bank did not blame AI for the rate rise. Energy prices drove the move to 4%; AI was one of the forces behind the activity strength that added to the case for tighter policy.

Key Takeaways

• The Bank of England's November Monetary Policy Report, published November 5 with an 8-1 vote to raise Bank Rate to 4%, estimates novel AI-related activity at just under a fifth of UK GDP growth since the start of 2026.
• The growth is upstream: development and implementation of AI technologies rather than adoption gains across the economy, a pattern the Bank calls consistent with a small productivity impact so far.
• Firms estimate AI lifted productivity by about 0.5% over three years, 0.1 to 0.2 points a year, while expecting 0.9 points a year over the next three; the Bank kept its 1% trend assumption and treats AI as an upside risk.
• Aggregate employment impact is close to zero so far, but AI adoption reduces hiring and replacement without raising firings, and some firms are at their headcount floor.
• AI hardware demand, memory chips above all, is pushing up Chinese and Asian export prices and feeding into UK import costs, an inflation channel the Bank tracks explicitly.
• Third-party context matches: Gartner found 3% of organizations meeting AI productivity targets, and the IMF has warned about the gap between AI spending and returns.

Sources: Bank of England, November 2026 Monetary Policy Report, report PDF, The Register, PPC Land, July 2026 Monetary Policy Report, Gartner survey via The Register, IMF warning via The Register