You don't have to leap to Industry 4.0

Gavin’s job title points to the cutting edge, but his instinct is to start nowhere near it. ‘This isn’t about looking at technology and going, how do I make that fit my business?’ he says. ‘It’s about looking at your business, your processes, your people, and identifying how to make those better.’

His starting point is what he calls baselining: ‘It’s trying to work out where you are now as a business. Not on digital transformation, not on digital thread, not on IoT – where are you as a business? Where are the bottlenecks? What does goods-in look like? How much money is tied up in inventory rather than R&D?’

Answer those first, and the right next move starts to reveal itself.

Crucially, that ‘right move’ depends entirely on what you make. Gavin draws a sharp line between two kinds of manufacturers. If you’re doing repetitive, high-volume work – Gavin uses the example of churning out 10,000 cans of baked beans a day – then technologies like machine vision or predictive maintenance can pay for themselves quickly. The data keeps coming, and the return is there. But if you build ten different products and make five of each in a month, the same technology won’t work for you because the return on investment won’t be there.

And the biggest mistake he sees? Trying to leap too far, too fast. Most UK manufacturers, he argues, aren’t at Industry 4.0 at all – they’re at Industry 3.0, and a jump straight to 4.0 is a genuinely risky move. The smarter play is to take incremental steps.

His favourite example costs almost nothing. Put a current clamp (an energy monitor) on a machine for a couple of hundred pounds, and you can work out what that machine costs to run, how hard it’s working, and how much spare capacity you actually have. This allows you to see how much of the day is spent making parts versus changing tooling. It’s a small, low-risk step that hands you real insight – and the confidence to make the next investment.

A current clamp (energy monitor) on a machine — Gavin’s favourite low-risk first step.

If there’s one thread Gavin kept returning to in our conversation, it’s people. And he’s blunt about the challenge. ‘We have an ageing workforce, and we have fewer people coming through than we have retiring,’ he says. ‘That is the reality.’ It’s a problem with no quick fix – he reckons solutions needed to start 20 years ago. But this problem also holds real opportunities.

The first is a long-overdue shift in attitude. When Gavin went to university, he says, the idea of an apprenticeship was ‘actively frowned upon’ – a wrong-headed way to look at how people learn. That narrative is finally changing, and with it comes a wave of younger people with what he calls digital-native skills. ‘These are people who grew up with an iPad, who grew up with a smartphone,’ he says. ‘They will look at your business differently to you, and that’s not a bad thing.’ They’ll bring different solutions to problems – if you let them into the conversation.

The flip side is the knowledge walking out of the door as that older generation retires: the tacit, hard-won know-how that’s notoriously difficult to hold on to. And this, Gavin says, is one of the areas where AI is genuinely earning its keep – capturing knowledge from experienced individuals, documenting it, and making it searchable so it can be passed on. It’s no substitute for having the person beside you, he’s quick to add. But the analogy he reaches for is a hopeful one: ‘If you think of us having a leaky bucket with twenty holes in it, it seems like we’re able to plug up at least half of those holes – which is better than where we were yesterday.’

Ten or fifteen years ago, Gavin says, the prevailing wisdom was simple: put a sensor on everything, read it in as fast as you can, and worry about what to do with it later. The industry was on track to generate eye-watering volumes of data, with little thought for the cost, the carbon, or the point of it all.

He has a sharper view now, and it’s probably the single most useful takeaway from the whole conversation.

Hoarding data you’ll never look at, he says, is ‘the equivalent of having a junk room for data. All you’re doing is taking up space.’ The value of data lies in the decisions it enables you to make. Real-time visibility only matters because it changes what you do next.

Get that right, and something rather lovely happens. Put the right data on a screen where people can see it – today’s throughput, say, machine one versus machine two – and you start to get a bit of healthy gamification, day shift competing with night shift to do more this week. But build a culture where spotting opportunities is rewarded, and your whole workforce effectively becomes a ‘digital suggestion box.’

Gavin’s example brings it to life. A building management system (BMS) already knows the temperature inside your building, and the outdoor temperature is free to pull in. Tie this together with your manufacturing resource planning (MRP), and someone on the floor might spot that you should schedule goods-in for, say, midday to 2pm – when the gap between inside and outside is smallest, so you lose the least heat opening the doors and spend less energy reheating afterwards. Nobody set out to find that saving – it surfaced because the data was visible and people were encouraged to look.

We couldn’t let Gavin go without asking about AI, though we kept the focus firmly on the factory floor rather than the chatbots. The space moves too fast to forecast far ahead, he says, so he’d only commit to where the real wins are landing now.

Top of the list is machine vision: cameras identifying defects, catching when something’s gone wrong, even checking whether people are wearing their PPE. Close behind is the speeding up of mundane work – filling out first-article inspections, or searching the vast standards documents and the how-to manuals for design software like NX or CATIA that engineers wade through, to find the exact thing you need for your exact situation, in seconds rather than hours.

He also offered a warning we think every manufacturer should hear. A lot of products coming to market ‘do a great job but are essentially just marketed really well’ – charging a premium for something you could often get cheaply, or free, if you knew where to look. The question worth asking, in his words, is simple: ‘Am I being sold snake oil?’ Having someone independent to test that judgement against is, he argues, exactly why places like the catapult centres matter.

Gavin is also clear about where AI can’t go yet – and the blocker isn’t the technology. A lot of what works today works precisely because you don’t need the model to show its working; you just take the efficiency and move on. But for anything safety-critical, or anything that has to be fully traceable, you have to be able to prove exactly how a decision was reached, and an AI model often can’t. ‘The technology might be there,’ he says, ‘but the regulations and the standards aren’t ready.’

We finished by asking Gavin what he’s most excited about, and his answer was immediate: supply chain integration.

Many of the big players in aerospace, defence, nuclear and green power have already invested heavily in their own transformations and are now hitting a ceiling on the returns. The next leap, he believes, comes from bringing their supply chains with them. ‘If they can get their supply chains to integrate the same way, a little bit more investment gets them about twice the reward.’ The hard part is the ‘how’ – creating data flows that move up and down the chain easily, and integrating with smaller firms that don’t have the budget for the same tools, licences or platforms. It’s a challenge of technology, yes – but also of standards, and of people, which is the theme of the whole conversation.