Six in ten retailers, wholesalers and manufacturers are worried about their supply chains and expect rising sourcing and transport costs over the next year. The right kind of AI could ease that pressure, according to experts. But adoption in these sectors lags behind the rest of the UK economy, and fewer businesses there say AI is actually improving how they operate.
AI use among UK businesses has climbed to 34.7% as of June 2026, according to Office for National Statistics data. Manufacturing sits at 26.9%, while the wholesale and retail sector sits at 28.4%. The industries that keep Britain’s shops stocked and factories running have adopted AI more slowly than the country as a whole.
Inefficient supply chains are one of several drivers of rising prices, and part of the cost tends to land on the shopper. AI built for forecasting could ease some of that pressure, according to Sigrún Gunnhildardóttir, CPO at AGR Inventory, a company providing demand forecasting and inventory planning software to wholesalers, retailers and manufacturers.
“A lot of it comes down to reactive procurement, paying a premium for expedited freight and emergency safety stock because the shortage wasn’t forecast. It’s a real driver of margin compression, and some of it shows up in the price consumers pay,” says Sigrún.
Nationally, close to 59% of businesses using AI say it’s improving their operations, the single most common reason given. In manufacturing, wholesale and retail, that figure drops to around 47%, roughly 12 points behind the national average, despite these being the sectors where operational tools like forecasting and inventory planning could counteract rising prices.
“Looking at the numbers, most of what’s been adopted in these sectors won’t help anyone decide how much stock to order,” says Sigrún.
Mostly free, mostly for content
Businesses in wholesale and retail are more likely than the national average to have adopted AI through free software, and manufacturers are noticeably less likely to buy dedicated tools. Among the AI these sectors do use, text and visual content creation are the most common applications. Data processing, the type of AI behind demand forecasting, trails well behind, according to the ONS figures.
Share of all businesses
| Uses AI | Text generation | Visual content | Data processing | |
| All UK businesses | 34.7% | 17.1% | 14.4% | 7.6% |
| Manufacturing | 26.9% | 7.5% | 14.6% | 6.1% |
| Wholesale and retail | 28.4% | 14.6% | 14.0% | 2.7% |
Share of AI-using businesses
| Adopted via free software | Adopted via purchased software | |
| All UK businesses | 44.5% | 42.8% |
| Manufacturing | 49.1% | 34.6% |
| Wholesale and retail | 53.6% | 41.3% |
Source: ONS Business Insights and Conditions Survey, Wave 159, 15 to 28 June 2026
“Off-the-shelf models have no visibility of stock positions, supplier lead times or demand history. Without that context, they can’t support a replenishment decision. And businesses need to be able to see what data an AI recommendation is based on, rather than simply being given an answer,” says Sigrún.
Businesses already know what they’re missing
Forecasting is the AI capability in highest demand among the manufacturers, wholesalers and retailers AGR works with, rated above every other option the company has asked about. The type of AI it relies on, advanced data processing and modeling, are among the least adopted in these sectors, according to the ONS figures.
“When our clients tell us what they want from AI, forecasting comes first. They want earlier warning on demand shifts, anomalies and the effect of promotions, so they can plan replenishment around it,” says Sigrún.
With six in ten of those concerned about their supply chains expecting sourcing and transport costs to rise over the next year, the gap between what these businesses want from AI and what they’ve adopted could become costly.
“These businesses want AI, and they’re under pressure on costs. But they don’t just want more AI. They want tools that understand their operational data, show them how recommendations are reached and leave them in control of the final decision. That’s very different from simply giving employees access to a general purpose AI tool,” says Sigrún.
Methodology
The sector-level figures on AI adoption, technology type, adoption method and purpose are taken from the industry breakdowns in the Office for National Statistics’ Business Insights and Conditions Survey (BICS), Wave 159 (15 to 28 June 2026).
Supply chain cost figures are drawn from BICS Wave 163 (17 to 30 August 2026) and are shares of businesses that reported concerns about their supply chains.
BICS excludes agriculture, energy, public administration, public health and education, and financial services, so its coverage reflects the UK’s private trading economy.
Customer insights are drawn from a survey of 50 members of AGR Inventory’s customer panel, conducted in July and August 2026 among retailers, wholesalers and manufacturers. More than half of respondents were supply chain decision makers. Respondents rated six AI-enabled capabilities for their value on a scale of 1 to 5.
