As factories across the UK open their doors for National Manufacturing Day, they do so at a point when British manufacturing is neither enjoying an uncomplicated recovery nor facing an uncomplicated decline. The latest Office for National Statistics figures show manufacturing output rose by 0.5% in the three months to July compared with the previous three months, while July itself delivered a 0.9% increase. Eight of the 13 manufacturing subsectors recorded monthly growth, led by computer, electronic and optical products.
Set against that improvement, however, is a much less comfortable operating picture in which stronger demand has yet to remove the pressures weighing on investment and recruitment. Make UK’s Q3 2026 Manufacturing Outlook reports improving orders and business confidence, but manufacturers remain cautious about taking on people as employment, energy and input costs continue to squeeze margins.
Even the scale of the sector can become obscured when the discussion revolves around short term output movements. Make UK’s Making More Than Products report puts UK manufacturing at £220 billion of economic output and 2.6 million jobs, with average manufacturing pay 8% above the whole economy average. National Manufacturing Day, now in its fifth year, is intended partly to make that industrial footprint more visible by bringing schools, colleges, jobseekers and local communities into factories that frequently bear little resemblance to the public image of manufacturing they are meant to challenge.
A recovery with qualifications
After several years in which manufacturers have had to absorb successive pressures rather than deal with them one at a time, even modest signs of recovery arrive with qualifications attached. Energy and employment costs remain high, trade conditions have become more complex and skills shortages persist, yet companies are simultaneously expected to automate, digitalise, decarbonise and invest in new production capacity.
Within July’s headline growth figure, the contrast between subsectors illustrates how uneven that recovery remains. According to the ONS, output in computer, electronic and optical products rose by 5.2% during the month, basic pharmaceutical products increased by 3.4% and basic metals by 2.8%. Rubber and plastics products and other non-metallic mineral products fell by 4.5%, while machinery and equipment declined by 1.5%.
Although Make UK’s latest outlook points to stronger demand and improving investment intentions, that has not yet translated into an equivalent appetite for recruitment or unfettered capital spending. Companies are still having to decide which investments can carry higher financing, labour and operating costs while delivering sufficiently rapid improvements in output or efficiency.
Complicating that calculation is the changing character of industrial investment itself, because buying a machine is increasingly only the beginning of the expenditure. Modern capital projects arrive with requirements around software, data, connectivity, cyber security, systems integration and workforce capability, all of which determine whether the equipment ultimately performs as intended. Installing technology may take months; embedding the knowledge required to use it effectively can take much longer.
It is this connection between capital and capability that Steve O’Keeffe, Regional Vice President, UK&I at Epicor, believes manufacturers cannot afford to overlook.
“Manufacturing has never been an industry afforded the luxury of standing still. As AI, robotics and automation become increasingly embedded in the sector, we need to ensure that the next generation is equipped with the skills needed to work alongside advanced technologies.”
For many manufacturers, therefore, the immediate question is no longer whether automation will become more deeply embedded in production, but whether the business has enough people able to specify, integrate, operate, maintain and improve increasingly sophisticated systems once they arrive.
Technology meets the factory floor
Nowhere is the distance between technological potential and operational reality more obvious than in artificial intelligence, where enthusiasm has grown considerably faster than deployment across core manufacturing processes. Make UK’s AI, Skills and the Future of the UK Manufacturing Sector report found that only 2% of manufacturers said AI was widely embedded across their operations. Fewer than 40% were using it in some areas, while more than half identified skills shortages as the main barrier to adoption.
Where AI has been introduced also matters, because adoption remains considerably more advanced in administrative functions than on the production line. The same research found operational adoption at 11% in production, 7% in supply chain and 6% in quality control, suggesting that much of industrial AI is still being tested around the edges of manufacturing rather than sitting at the centre of production control.
That gap between interest and implementation was evident at a National Manufacturing Day event hosted by Safran Nacelles in Burnley and organised by RTC North, Made Smarter North West and Burnley Council. Almost 50 delegates from sectors including aerospace, automotive, machinery, plastics, textiles and printing discussed applications already saving time and cost, but their experience also exposed the less glamorous obstacles of limited specialist skills, poor data, security concerns, lack of time and growing AI fatigue.
At Safran Nacelles, where AI is being explored alongside manufacturing execution systems, connected machinery, simulation and automated inspection, the technical discussion quickly comes back to the quality of the information on which those systems depend. Michelle Hubel, Head of Digital Transformation, described the issue succinctly: “We all think we’ve got good data until we actually start looking at it.”
Behind that observation sits a familiar industrial problem. Few manufacturers begin a digital project with a perfectly structured data estate, unlimited integration capacity and a team of specialists waiting to build models around it; they begin with installed machinery, legacy systems, production commitments and employees whose first responsibility is still to keep output moving.
Rather than beginning with the technology and searching for somewhere to deploy it, Kevin Smith, Lead Technology Adoption Specialist for Made Smarter North West, argues that manufacturers should work back from the problems consuming time and resource.
“AI doesn’t automate jobs, it automates tasks. Look at what takes up people’s time, what frustrates them and where AI could make a difference.
“Then start small, get some quick wins and build confidence and skills. Keep people at the heart of it and make sure you’re solving a genuine business problem. The opportunity is huge, but ultimately AI has to deliver value for your business.”
That distinction may determine how much of the current experimentation survives into routine production. An inefficient process remains inefficient when wrapped in newer technology, while poor data does not become more reliable simply because an algorithm can process it faster. For many factories, the difficult work lies in joining new tools to existing equipment, processes and people without adding another layer of complexity that produces impressive demonstrations but little measurable improvement.
The skills problem is broader than recruitment
Running beneath those questions of automation, data and AI is a workforce problem that has become broader than the traditional shortage of engineers and technicians. Make UK’s Shape of British Industry research found that 99% of manufacturers believe access to skills will shape future growth plans, while half identify skills shortages as their biggest barrier to growth. At the same time, 65% plan major investments in digitalisation and AI.
Recruitment will remain part of the answer, but relying on the external labour market to provide every capability manufacturers need becomes less realistic as technical requirements change. Internal progression, retraining and the formal recognition of knowledge already built through years on the factory floor consequently become more important, particularly for SMEs competing for people with larger employers.
Industrial packaging manufacturer WERIT UK provides a useful example of what that can mean in practice, having invested £8 million in its Irlam manufacturing facility while doubling blow moulding capacity from three machines to six. Alongside the capital programme, the company has developed an engineering apprenticeship pathway supported by internal progression, cross training and technical training.
The experiences of two employees completing Level 3 engineering apprenticeships this year demonstrate how widely those development routes can differ. Alex Davies, 23, joined WERIT in production around five years ago and progressed through a supervisory role before moving into maintenance engineering, completing an electrical and electromechanical qualification while working full time. His work now involves maintaining and repairing machinery across the manufacturing operation.
Having arrived at the same broad engineering function by a very different route, Tomasz Kreft shows why existing experience can be as important to the skills equation as attracting new entrants. Now 47, he joined WERIT as a production operator in 2008 after moving to the UK from Poland, progressed to shift leader, moved into engineering and is now engineering manager. After years of practical experience spanning welding, CNC machinery and engineering, he has completed a Level 3 Science Industries Maintenance Technician Apprenticeship with Distinction.
For Sunny Prakash, Managing Director of WERIT UK, the two careers demonstrate why manufacturers cannot treat recruitment as the only route to increasing technical capability.
“The manufacturing skills challenge will not be solved by recruitment alone. SMEs need to recognise the potential within their existing workforce and give people opportunities to develop.”
Seen in that context, apprenticeships become more than an entry route for school and college leavers. They can also support career changes, internal progression and recognised qualifications for employees who already possess years of practical knowledge, while allowing employers to develop capability around the equipment and processes they actually use.
Nor does increasing automation reduce the importance of that expertise in any straightforward way. As WERIT’s investment illustrates, additional production equipment creates more assets to commission, monitor, diagnose, repair and improve, meaning the value of employees who understand both the process and the machinery can rise as the factory becomes more technically capable.
A similar conclusion emerges from companies approaching the issue through software rather than machinery. Ronan Hafferty, founder of Hafferty Interiors, has used Made Smarter North West support to develop an AI strategy and explore applications across design and manufacturing, but his experience has reinforced the need for judgement rather than removed it.
“AI has already made a real difference, particularly in design where it has saved us a huge amount of time and cost, and it is blowing our minds every week with how quickly it is developing. But you can’t be lazy with it. You can’t get rid of common sense and the human element.”
As production systems become more connected, that combination of digital capability and accumulated manufacturing knowledge is likely to become more valuable. Software can identify patterns, automation can remove repetitive tasks and better information can expose waste, but deciding whether an intervention makes engineering and commercial sense still depends on people who understand what is happening beyond the screen.
From investment to productivity
Once the excitement around pilots, demonstrations and technology announcements begins to settle, the next phase of this investment cycle will be judged by rather more conventional industrial measures. Whether AI improves throughput, whether automation raises uptime, whether connected systems reduce scrap and whether better data shortens lead times will matter considerably more than the number of manufacturers able to say they are experimenting with the latest technology.
Getting from one to the other places greater emphasis on execution, because the difficult decisions begin after the technology has been selected. Manufacturers need to determine which constraint is actually worth automating, whether the data feeding a system can be trusted, how new equipment will connect with the rest of the process and who will retain the knowledge needed to maintain and improve it after the implementation team has gone. A pilot that remains permanently a pilot may demonstrate technical possibility, but it does not improve factory productivity.
The same logic applies to workforce investment, particularly when recruitment remains difficult and technical capability is increasingly specific to individual processes and equipment. Developing an experienced operator into a maintenance technician, teaching engineers to work with connected systems or giving production teams the ability to interrogate operational data may produce a more immediate return than competing indefinitely for an ideal candidate whose skills are already scarce across the market.
With manufacturers still protecting margins against high employment, energy and input costs, the threshold for approving both capital and training expenditure is likely to remain demanding. That does not remove the need to invest; it simply makes vague promises of transformation less persuasive when compared with projects that can demonstrate a credible route to increased output, lower waste, better reliability or more effective use of labour.
There may be an advantage in that greater scrutiny. AI, automation and connected manufacturing do not need to revolutionise every process in order to justify their place on the factory floor, provided they solve enough persistent and expensive problems to earn the investment. Manufacturers that can identify those problems accurately, develop the people capable of addressing them and then carry successful applications from trial into production will be better placed than those accumulating technology without the organisational capability to exploit it.
As UK manufacturing moves through the final months of 2026, the distinction between possessing technology and possessing industrial capability is becoming increasingly important. There is no shortage of machinery, software or digital tools available to buy, and no shortage of productivity problems to which they might be applied. The harder task is connecting capital, data and engineering knowledge closely enough that improvements survive beyond the pilot project and become part of everyday production.
