Two manufacturers, two different realities

Walk into two factories of similar size and you will often find them operating in different worlds. One runs on connected systems where production, stock, purchasing, and finance share a single source of truth. The other reconciles yesterday’s output from a spreadsheet, checks stock on a system that does not talk to the factory floor, and waits on end-of-week reports assembled by hand from three sources.

Nothing has broken. But the second firm carries an enormous structural cost: wasted time, delayed decisions, an inability to quote accurately, and no way to deploy the AI and analytics tools that could transform its productivity. Competitors with connected infrastructure pull further ahead each quarter.

What the research found

The vendor data myth

Vendor surveys and official government research produce dramatically different pictures. AWS reports that 64% of UK organisations use AI. DSIT, built on 3,500 structured interviews across the full business population, puts it at 16%. That gap is not noise. It is the distance between what large, well-resourced enterprises are doing and what the SME manufacturers who make up most of both countries’ industrial base are actually doing.

The Cambridge Bennett School data confirms it: large UK firms nearly doubled their AI adoption to 44% between 2023 and 2025, while small firms reached just 26%. An SME reading vendor headlines and concluding it is too late or too expensive to catch up is drawing the wrong conclusion from the wrong data. Most of its direct competitors are at roughly the same stage.

The cost of staying disconnected

Disconnection is not a neutral state. It carries an ongoing cost in wasted admin hours, delayed decisions, inventory errors, and missed opportunities. Manufacturers lose an average of 45 hours per month to debugging caused by legacy industrial code tools (Copia, 2025). Irish SMEs without ERP face an estimated EUR 47,000 in annual cost from inefficiency and missed opportunity (ProfileTree, 2025). Across the wider economy, legacy systems are estimated to cost the UK around GBP 45bn a year in lost productivity (Stromasys, 2026).

The most consequential hidden cost, though, is the AI lockout. Predictive maintenance, demand forecasting, and quality analytics all require clean, structured, integrated data. Legacy systems hold data in disconnected silos, which structurally prevents these tools from being deployed at all.

A five-phase roadmap to close the divide

Most firms see measurable improvements in management visibility and efficiency within Phase 1. There is no requirement to reach Phase 5 before realising value. The important thing is to start with an honest assessment of where current systems are generating the highest cost.