Why Britain's Industrial Strategy Needs Regulatory Infrastructure

Britain has world-class engineers. The constraint that is becoming national is not talent but the speed and certainty with which regulated engineering knowledge can be found, applied and trusted.

Executive summary

Britain's debate about advanced manufacturing tends to focus on talent and capital. Both matter. But in regulated sectors, the binding constraint is increasingly something less discussed: the time and certainty with which an engineering team can determine which rules apply to a product, what evidence is needed, and whether a change is compliant.

That capability is not regulation itself. It is the tooling, data and process that let engineers act on regulation quickly and reliably. It behaves like infrastructure. When it is slow and manual, every product and every change re-incurs the same cost of finding and applying the relevant knowledge, and small firms carry that cost least efficiently.

This article argues that regulatory knowledge, rendered into usable digital engineering infrastructure, is a national productivity question for advanced manufacturing. It looks at where the constraint bites across aerospace, medical devices, automotive, defence, electronics and semiconductors, why knowledge rather than talent is the limiting factor, how it connects to supply-chain resilience, and what the digital engineering layer that addresses it actually consists of. It avoids policy advocacy and keeps to the structure of the problem.

A constraint that is shifting

For two decades the standard account of Britain's industrial position has centred on people and money: enough skilled engineers, enough patient capital, enough demand. These remain real questions.

There is a third constraint that is becoming more visible, particularly in the regulated parts of the economy. It is the speed at which an engineering organisation can answer a specific class of question: for this product, in these markets, which rules apply, what evidence must be produced, and what does a given design change do to all of that.

In an unregulated software product, that question barely exists. In a regulated hardware product, it is on the critical path to market, and it recurs every time the design changes.

Talent helps answer the question, but it does not remove it. The knowledge a specific product needs is distributed across standards documents, regulator guidance, historic test evidence and the memory of experienced individuals. It does not scale cleanly with headcount. Hiring another engineer does not make the standards shorter or the cross-references faster to trace.

This is why it is useful to think of the capability as infrastructure rather than as a skills gap.

What "regulatory infrastructure" means

Regulatory infrastructure is not the same as regulation. Regulation is the rules. Regulatory infrastructure is what lets an engineer act on the rules quickly, consistently and with an auditable trail.

The analogy is to other kinds of infrastructure. Physical infrastructure, such as roads and power, does not produce goods; it lowers the cost of producing and moving them. Digital infrastructure, such as cloud platforms and continuous integration, does not write software; it lowers the cost and raises the reliability of shipping it.

Regulatory infrastructure, in the same sense, does not certify anything and does not replace a regulator. It lowers the cost and raises the reliability of determining what compliance requires and whether a design meets it. Today that determination is largely a manual act: a qualified person reads documents and applies judgement. The infrastructure question is how much of the finding, mapping and checking around that judgement can be made fast, consistent and traceable.

When this layer is weak, the symptoms are familiar. Compliance knowledge lives in a few people's heads. The same interpretation is redone for each new product. The link between a requirement and the evidence that satisfies it exists only in someone's filing system. A change late in a programme triggers a slow, expensive scramble to work out whether it matters.

Where the constraint bites

The regulated sectors most associated with UK advanced manufacturing each carry a large compliance surface, and in each the knowledge to navigate it is scarce and slow to apply.

Aerospace. Civil aviation design and production are regulated in the UK by the Civil Aviation Authority, and across the EU by the European Union Aviation Safety Agency. Design approvals, continued airworthiness and a dense body of certification specifications mean that even contained changes can carry substantial evidence requirements. The knowledge to scope that work is held by a relatively small population of specialists.

Medical devices. Devices placed on the UK market are regulated by the Medicines and Healthcare products Regulatory Agency, and in the EU under Regulation (EU) 2017/745, the Medical Device Regulation. The regime ties classification, evidence, clinical data and post-market obligations together, and the work of determining what applies to a given device and change is itself a specialist task.

Automotive. Vehicle type approval draws on the United Nations regulations administered under the UNECE 1958 Agreement, alongside national arrangements. The number of applicable regulations for a single vehicle programme is large, and changes propagate across them in ways that require careful tracing.

Defence. UK defence procurement uses Defence Standards (DEF STAN) and a layered assurance regime. The compliance knowledge is again concentrated and document-heavy.

Electronics and radio. Connected products sit on the surface described in the companion article on component changes: the EU Radio Equipment Directive (2014/53/EU) and its harmonised standards, UKCA marking, and, for export, regimes such as the FCC rules in the United States. A single device can face several jurisdictions at once.

Semiconductors. The UK set out its approach to this sector in the National Semiconductor Strategy published in 2023. The sector's design and manufacturing activity sits upstream of all of the above, and its products inherit and impose compliance requirements across the chain.

The common pattern across these sectors is not a shortage of capable engineers. It is that the regulatory knowledge each product needs is large, distributed, and expensive to apply, and that this cost is paid again for every product and every significant change.

Why knowledge, not talent, is the bottleneck

Talent is, in principle, a solvable problem. People can be trained, recruited and retained, and the UK has a strong base of engineering education and skill. The number of competent engineers is a constraint that responds to investment over time.

Regulatory knowledge access behaves differently. Consider what happens when a regulated-hardware firm grows or takes on a new product.

The knowledge needed is specific to the product, the markets and the applicable standards, and it has to be reassembled each time from documents and individuals. Adding engineers does not reduce the length of the standards, the number of cross-references, or the time to trace a change through them. It distributes the same slow work across more people.

The knowledge is held unevenly. A handful of experienced individuals often carry the practical map of what applies and why. When they are busy, or leave, the organisation's effective certification capacity drops sharply, regardless of total headcount.

The cost is re-incurred per change. Because the link between a change and its compliance impact is rarely written down in a reusable form, each change starts the analysis close to the beginning. There is little compounding. The hundredth change is not much cheaper to assess than the tenth.

Small and medium-sized firms feel this most acutely. A large prime can maintain a substantial compliance function. An SME cannot, so the same regulatory surface consumes a much larger share of its engineering capacity, and the founder or lead engineer often absorbs the work personally. For a country that wants more small advanced-manufacturing firms to scale, this is a structural drag that talent policy alone does not address.

Note for the verified edit. Quantitative evidence on compliance and certification effort as a share of engineering time in regulated UK sectors would strengthen this section. I have not estimated it. With web search enabled, defensible figures from primary or reputable sources can be sourced and inserted.

The policy backdrop, factually

UK industrial policy has, in recent years, named advanced manufacturing and critical technologies as priorities. The relevant documents are worth reading in the original.

The UK Science and Technology Framework, published in 2023, set out the government's approach to a set of critical technologies. The National Semiconductor Strategy addressed that sector specifically. The Advanced Manufacturing Plan, published in 2023, addressed manufacturing more broadly. A modern industrial strategy was set out for consultation in the green paper Invest 2035 in 2024; the current status of that strategy should be checked on gov.uk, since these documents have been revised and superseded over time.

The institutions that underpin standards and measurement in the UK are long-standing. The British Standards Institution is the national standards body. The National Physical Laboratory is the national metrology institute. Both are part of the existing infrastructure on which any digital regulatory layer would build.

The point here is not to advocate a particular policy or spending level. It is to observe that the strategic priorities already named, advanced manufacturing and critical technologies, depend in practice on engineering teams being able to act on regulation quickly. The enabling layer that makes that possible is a complement to talent and capital, not a substitute for either.

The supply-chain resilience angle

Regulatory infrastructure connects directly to a goal that policy already cares about: supply-chain resilience.

Resilience is often framed as having alternative sources for critical components. That is necessary, but it is incomplete. As the companion article on component changes sets out, switching to an alternative source is not free in a regulated product. A second source can change parameters that feed an applicable limit, which can require re-evaluation, re-testing and updated documentation.

So the practical ability to use an alternative supplier depends on the ability to re-qualify quickly. A firm that can determine the certification impact of a substitution in days can respond to a disruption. A firm for which that determination takes weeks of specialist time cannot respond at the speed a disruption demands, and may carry the risk of either a stranded design or a non-compliant shipment.

In that sense, the speed of regulatory knowledge access is part of supply-chain resilience, not separate from it. Stockpiles and second sources reduce exposure to disruption; fast, reliable re-qualification reduces the cost of acting when disruption arrives.

Engineering productivity in regulated sectors

Productivity in software engineering has improved through infrastructure: version control, continuous integration, automated testing, deployment pipelines. The work that used to be slow and manual was made fast and repeatable, and the engineer's time moved to higher-value problems.

Regulated hardware engineering has had a parallel set of improvements in design and simulation, but the compliance determination has lagged. The productivity of a regulated-hardware engineer is gated, in part, by how quickly they can answer compliance questions, and that step is still largely manual.

This is where digital engineering infrastructure matters for national productivity. The gains in this domain do not come mainly from designing faster; they come from closing the gap between making a change and understanding what the change costs in compliance terms. That is a tooling and data problem, and it is addressable.

What the digital engineering layer consists of

A practical digital engineering layer for regulated work has a few components, each of which is an active area of development rather than science fiction.

Structured representation of standards. The content of a standard, its definitions, parameters, limits and the conditions under which they apply, represented as data that software can check against rather than only as prose to be read. International standards bodies have programmes working towards machine-readable standards, and there are existing examples in adjacent domains, discussed in the companion article on executable standards.

Traceability from requirement to evidence. An explicit, machine-navigable link from a requirement, to the clause that imposes it, to the evidence that satisfies it, for a specific configuration. This is the connective tissue that turns a pile of documents into something queryable.

Deterministic verification. A checking step that, given a configuration and the applicable structured rules, produces a result that is traceable to the clause and the evidence, and that is reproducible. This does not interpret ambiguity or replace a body; it mechanises the parts that are mechanical, and leaves judgement where judgement belongs.

These are the same building blocks that any serious attempt at regulatory infrastructure has to assemble. They are described in more depth in the companion article on why standards should be executable.

How Crado approaches this

Crado builds engineering decision infrastructure for regulated hardware. It creates a deterministic chain from an engineering change through to applicable clauses, required evidence and a verifiable decision, so that the certification impact of a change can be examined before a product reaches the test lab.

AI is used only to read and structure documents such as standards and test reports. The engineering decision comes from deterministic verification against recorded rules, and every result is traceable to the specific clause and the specific evidence it rests on.

Crado does not certify products, does not guarantee compliance, and does not replace engineers, accredited laboratories or certification bodies. The aim is to make regulated engineering knowledge faster to apply and auditable when it is, which is the practical content of regulatory infrastructure for a single firm.

Conclusion

Talent and capital are necessary conditions for a strong advanced-manufacturing base. They are not sufficient on their own in regulated sectors, where the speed and certainty of acting on regulation is itself a constraint on how fast firms can build, change and scale.

That constraint is a knowledge-access problem, and knowledge-access problems are addressed with infrastructure. Rendering regulatory knowledge into fast, consistent, auditable digital engineering infrastructure is a productivity question for the sectors the UK has already said it wants to grow. It is a layer that makes existing talent and existing capital go further, which is usually where durable national advantage comes from.

1. UK Science and Technology Framework (2023) — GOV.UK
2. National Semiconductor Strategy (2023) — GOV.UK
3. Advanced Manufacturing Plan (2023) — GOV.UK
4. Invest 2035: the UK's modern industrial strategy (green paper, 2024; verify current status) — GOV.UK
5. Civil Aviation Authority — Civil Aviation Authority
6. European Union Aviation Safety Agency — EASA
7. Medicines and Healthcare products Regulatory Agency — GOV.UK
8. Regulation (EU) 2017/745, Medical Device Regulation — EUR-Lex
9. UNECE vehicle regulations (1958 Agreement) — UNECE
10. Radio Equipment Directive (2014/53/EU) — EUR-Lex