AI in the back office. Engineering IP in the prompt.
Most manufacturers use AI somewhere, almost all of it in the back office. Meanwhile drawings, formulations and pricing get pasted into tools nobody has reviewed, just as customers begin asking suppliers about cyber and AI controls.
of manufacturers have AI widely embedded
Make UK, 2026of manufacturers' AI use is back-office
Make UK, 2026adoption in transport and storage; 90% neither use nor plan to
DSIT, 2026No manufacturing AI rule book. The rules you already follow still apply.
UK GDPR and the Data (Use and Access) Act 2025 apply to staff and customer data in any AI tool, and the Information Commissioner's Office (ICO) published findings on AI in recruitment in 2026 that apply to any employer screening candidates. The National Cyber Security Centre (NCSC) recommends that large buyers require Cyber Essentials, its baseline scheme, from suppliers, so customers increasingly ask for it. Rules on export control and health and safety reporting continue to apply to how information is handled and recorded.
Summarised from ICO and NCSC publications and law-firm commentary. The Make UK figures and the claim that public bodies must ask suppliers to declare AI use (Procurement Policy Note 017) are carried over and need checking at source. The NCSC playbook is guidance, not law. Not legal advice.
Six problems that keep coming back.
Drawn from surveys of manufacturers and supply-chain guidance.
Drawings, bills of materials, formulations and test results are the value of the business. Pasted into a free assistant, they leave your control.
Customer specifications and pricing go into AI tools to speed up quotes, often against the confidentiality terms of the contract.
Planning, quality and maintenance software now ships AI features. They are switched on by default and rarely appear on any list.
Large customers ask suppliers for Cyber Essentials and evidence of data controls. AI tools add new data flows to those answers.
Plant and office staff are busy. A policy no one has time to read changes nothing, and nobody owns the follow-up.
Older machines and control systems were not built to be exposed. Connecting them to cloud AI without a plan raises risk beyond data protection.
A leaked specification is not recoverable.
A competitor who sees your formulation, or a customer who learns their drawings were in a public tool, does not forget. The loss may be a contract, an exclusivity clause or a supplier-approval status. Compared with this, the cost of knowing which tools are in use is small.
Start with the list and three rules.
Smaller manufacturers do not need an AI governance team. They need managed accounts for the tools staff already use, a register that includes AI inside operational software, and three clear rules: what may be shared, who approves, and where AI is not used. That is enough to answer a customer questionnaire and to protect the intellectual property that matters.
- Quotes, tenders and technical documentation
- Supplier and customer communications
- Quality and maintenance notes
- Planning and scheduling support
- Drawings, formulations and pricing in public tools
- AI features inside planning and quality software nobody reviewed
- Skills gap: staff lack time to train
GenWA redacts intellectual property and pricing on connected tools, keeps a register of AI inside operational software, and applies rules on what may be shared and by whom. A shared Library for quoting and documentation is being built next and is not live.
Four agents, trained for this sector.
The rules, as controls.
- Cyber Essentials and NCSC guidance
- Export control and customer confidentiality terms
- Health and safety reporting records
- UK GDPR and the Data (Use and Access) Act 2025
- ISO 9001 quality management records
- Customer supplier-assurance questionnaires
Built around owner-managed plants and multi-site groups.
Manufacturers range from family-owned plants to multi-site groups. We shape each route to yours.
An independent audit with first findings in 48 hours and a plain 90-day plan for the managing director and operations director.
Start the health checkAdvice: policy and shop-floor trainingA one-page AI policy and short training for office and plant staff, so rules are understood and followed.
See advisoryProduct: control toolkitRegister, Guard rules, log and Evidence Packs, hosted in your AWS or Azure account, or GenWA AWS as agreed.
Find the right productProduct: agents at workQuote drafting, supplier communications and invoice processing with named owners.
Meet the teamIncubator: one function, twelve weeksWe define, build and run one controlled function, such as quoting support. You own the intellectual property, prompts and record.
See the incubatorAccelerator: for industrial technology foundersBy application only, ten places a year. Senior AI leadership time, advisory and platform access, paid for by equity, project, retainer or licence.
See the acceleratorWhat Guard and Evidence typically hold for this sector.
Control-mapped, not certified. We prepare the evidence; accredited bodies issue certification.
The UK government-backed baseline cyber scheme. We map AI controls alongside it and prepare evidence; accredited bodies certify.
The information security standard that larger customers may ask for, with AI controls mapped.
The international standard for managing AI. Readiness mapping.
Staff and customer data in AI tools, with redaction and review evidence.
Your contractual and export restrictions, loaded as rules on what may be shared.
Where products or AI features reach EU customers.
Three role holders, three different problems.
Carries the contract and IP risk. GenWA gives a one-page view of AI use, three clear rules and evidence for the next customer audit.
Wants quotes, planning and supplier messages faster without errors. GenWA puts named owners and approval steps on those workflows.
Has no time for another project. GenWA is set up with you, shows AI inside systems already in use and keeps the register current.
From this sector.
What does GenWA do for manufacturers and logistics firms using AI?
Redaction for intellectual property and pricing on connected tools, a register for AI inside operational software, and rules on what may be shared and by whom. The audit shows what else is in use.
Is AI governance too much for a firm without an IT team?
No. A small firm can begin with managed accounts, three rules and a one-page monthly view. The SME audit fee is indicative at £7,500 excluding VAT for 50 to 250 people, credited in full against year one of a platform plan.
Our planning and quality software suppliers say their AI is safe. Do we still need a register?
The register lists AI features inside operational software with an owner and a data classification, so you decide with facts. We do not assess a supplier's own claims.
What does Cyber Essentials have to do with AI?
Customers often ask suppliers about cyber controls, and AI tools add data flows. The NCSC recommends large buyers require Cyber Essentials from suppliers. We map AI controls alongside it and prepare evidence; accredited bodies certify.
How widely is AI embedded in UK manufacturing?
Make UK's 2026 research found 83% of manufacturers' AI use is back-office and only 2% have AI widely embedded. DSIT's 2026 data show 10% adoption in transport and storage. Check the sources before quoting. A register is a sensible first step.
Can AI stay out of the plant floor?
Yes. Rules can name where AI must not be used, such as safety-critical settings, and the register shows if a supplier switches a feature on.
How does a manufacturer start?
Take the AI health check. A usage-export check runs locally in your browser. First findings arrive within 48 hours and the audit with a 90-day plan within ten working days.
Do you know where your drawings and prices have been pasted?
First findings within 48 hours of the health check. Audit fee credited in full against year one of a platform plan.