Unlocking the next wave of robotic automation

AROUND 542,000 industrial robots were installed globally in 2024, according to the latest World Robotics 2025 report from the International Federation of Robotics (IFR). This resulted in a global density of 177 robots per 10,000 manufacturing employees. In contrast, public reporting based on the IFR report pegs the UK at roughly 104 robots per 10,000 manufacturing workers and about 24th globally, well below Western Europe’s 267. Meanwhile, the number of new UK installations fell 35% to just 2,500 units in 2024. This followed a peak in 2023, just before the end of the government’s popular super-deduction tax break.

This matters because research from trade association TechUK suggests that British manufacturers could enjoy a 22% increase in productivity if their automation levels matched those of the world's most automated countries. Clearly, many companies need sustainable incentives to turn their awareness of automation into firm investment decisions.  

Driving change

On the positive side, the British government seems committed to modernisation, with the Modern Industrial Strategy and Advanced Manufacturing Plan dedicating roughly £2.8 billion toward R&D in automation, robotics, and smart factories. More specifically, the Robotics Adoption Programme is a five-year, £52 million investment running through 2030 led by UKRI and Innovate UK to help organisations transition from testing to active deployment. Robotics Adoption Hubs are also worth a mention: This regional network provides vendor-neutral proving grounds, complete with shared testbeds, expert integration advice, and links to tech vendors and finance.

In addition to these ‘carrots’, industry also faces significant ‘sticks’ in the form of challenges that highlight an urgent need modernise.

First, consider the squeeze on labour. Make UK’s Industrial Strategy Skills Commission report 2025 estimates that the currently 55,000 unfilled manufacturing vacancies cost around £6.0 billion in lost output per year. That’s about £16.5 million a day. This problem impacts the jobs market from low-skilled roles such as packing through to skilled professions such as welding.

Next, look at how far international competitors are racing ahead while British firms hesitate. China reached 567 robots per 10,000 manufacturing employees in 2024, ahead of Germany (449) and Japan (446), according to IFR.

What is holding UK-based manufacturers back? Upfront costs are one factor, but the barriers extend beyond that. Adopting robotics is not just about buying off-the-shelf, plug-in solutions. It also requires strategic planning, workforce training, and operational adjustments. Many businesses struggle with the resulting uncertainties and the practical challenges of integrating robotics into their processes.

Ironically, the skills shortage is part of the problem, with nearly half of manufacturers (46%) citing a lack of technical expertise as their primary hurdle to deploying advanced equipment, according to figures fromMake UK. Many companies are also wrestling with legacy IT and complex factory floor set-ups, making the prospect of having to integrate new and existing systems a deterrent. Then there’s the fact that many SMEs delay implementation because they struggle to accurately measure the operational return on investment (ROI) they can expect from complex automation.

Inflection point

While these challenges have proved intractable in the past, the arrival of AI and other recent developments means UK industry is finally in a strong position to invest and compete more effectively through automation. That’s because AI is toppling all the traditional barriers.

Unprecedented flexibility and intuitive ease-of-use mean that discussions around AI are no longer focused on some vague future promise, but on actual proven benefits that robot users around the world are already enjoying.

See, simulate, adapt

The most widespread applications within robotics generally deploy AI in one or more of three specific areas.

The first is machine vision. Traditional machine vision might recognise an object presented in a consistent orientation but struggle to interpret an alternative object or one arriving at the ‘wrong’ angle. The addition of AI means robots can now perform random bin picking, depalletise mixed loads, carry out visual inspections for quality assurance, and adapt automatically to handle seasonal variations or random packages, all without the need for human intervention or retraining.

Commercially, it means robots can now be deployed in high-mix/low-volume environments where conventional fixed automation would be difficult to justify.  For instance, the latest AI-powered order fulfilment solutions can typically handle up to 1400 picks per hour during peak periods at greater than 99.5% picking accuracy. Even in tricky applications such as fashion distribution, an AI-vision-equipped robot can be trained to spot stable grasp points on highly deformable garment pouches (polybags), solving issues of light reflection and pouch sagging to deliver up to 1300 picks per hour.

Ultimately, the performance in individual applications will depend on factors such as the available training data, lighting, edge cases, and safety validation. It would be overstating it to suggest that these are plug-and-play solutions that will deliver peak performance ‘out of the box’ without laying the proper groundwork.

The second arena for AI in robotics is the use of simulation tools to de-risk investments and optimise robotic systems from day one. Digital twins and similar models can now achieve a level of hyperreality that enables installations to be designed, simulated, programmed, and stress-tested in a virtual environment ahead of physical installation.

Predictable outcomes strip much of the risk out of investment and deployment, thus bolstering the business case for robotics. Virtual modelling results in fewer surprises, less disruption, faster commissioning, a clearer view of throughput, and easier operator training. Users can be sure that a robot will be able to reach every part of a pallet without hitting a support beam, for example, or understand exactly where a robot arm will stop if it breaches a light curtain.

Accurate modelling also boosts productivity. Virtual commissioning can be used to analyse cycle times down to the millisecond, optimising paths and motion profiles to ensure the system delivers the required throughput and ROI.

The third area where AI excels is in helping robots achieve autonomous decision-making. A similar argument applies in any process that requires on-board intelligence, but the benefits of autonomy are perhaps easiest to understand in the case of autonomous mobile robots (AMRs). AMRs are increasingly deployed to provide intralogistics, line-side replenishment and flexible routing between production cells in dynamic environments such as factories and warehouses. They don’t need tracks or other infrastructure and can even navigate safely around people in a shared space.  

ABB’s AMR fleet can deliver a 90% enhancement in overall logistics efficiency, provide a 25-35% increase in production capacity, and a 20% reduction in setup or commissioning time, for example.

From concept to reality

This all sounds great in principle, but many companies still need support to take their first steps in robotic automation. Reputable suppliers can help spot the unique features in each potential deployment, but the shape of underlying roadmap is pretty consistent.

First, map out a constrained process where it’s clear that improvements can be made. Measure everything, such as baseline labour, quality and/or downtime limitations with the current set-up. Don’t forget to ask operators about their biggest bugbears. Once you’ve defined the requirements, work with the supplier to simulate the proposed solution. Today’s advanced virtual models should help users quantify the expected benefits and avoid potential pitfalls. Next, validate safety and cybersecurity provisions before deploying the first system. Measure the performance of this pilot and check it’s delivering as expected. If so, the user now has a set of reusable specifications and standards they can scale up for use in future automated systems.

In terms of finance, there may be government support available to pump prime investment, and solution providers may offer favourable terms to make systems more affordable.

Stakeholder involvement

AI has already started to catalyse investment in robotic automation but buy-in from key stakeholders will be needed to embed it long-term within the wider manufacturing landscape.

Companies should recognise that robots can augment human capabilities and allow businesses to use their workforce to the best effect. Employees should be reassured that automation is primarily about reskilling, not job losses. And for government, reinvigorating the UK’s industrial performance requires consistent support to promote investment. Financial help is always welcome, but support from other areas such as education will also be needed. 

No one is arguing that the UK should automate for automation’s sake, but the right systems can plug the productivity gap and bridge the skills divide to build flexibility and resilience. Look around and almost every manufacturing business will certainly be able to find at least one bottleneck that could be improved. Recognising that makes it easier to take the first step.

Julian Ware is UK and Ireland sales manager at ABB Robotics UK

www.abb.com/global/en/areas/robotics