AI in practice: key takeaways from the Digital Champions Forum

Attendees, including Siemens Digital Industries, Ricoman, Amtrim, Agreka Build, BEP Surface Technologies, Firstplay Dietary Foods, Velden Engineering, Bindatex and Additive Manufacturing Solutions, heard insights on identifying practical AI opportunities, understanding its limitations and delivering measurable value.

Kevin Smith, Lead Technology Adviser at Made Smarter North West, said the best place for most SME manufacturers to start is with contained IT applications that are lower risk, easier to test and can demonstrate value quickly.

That might mean reducing administration, improving reporting, supporting quoting or planning, or handling customer communications. But his advice was to “start with tasks, not technology”, identifying where AI could deliver value and whether the data, process and level of risk make it feasible.

Kevin said these early applications should be treated as capability-building pilots, not isolated experiments. As well as delivering efficiency gains, they should build AI literacy, governance, process discipline and confidence.

Capturing that learning is crucial, Kevin warned, otherwise businesses risk accumulating disconnected tools rather than using each pilot to build a wider approach to AI adoption.

While IT AI offers a practical starting point, Kevin stressed it shouldn't be the end goal. Manufacturers can pursue quick wins while exploring opportunities closer to production, where AI could deliver greater operational value.

Applications include predictive maintenance, machine vision, condition monitoring and process control, helping manufacturers anticipate equipment failure, detect quality issues, improve processes and optimise resources.

But closer to production, greater assurance is needed. “The question is no longer whether it works.” Manufacturers need to know AI can be trusted in live conditions, with reliable data, effective integration, proper validation and clear ownership.

This reflects the approach set out in Made Smarter's AI Adoption in Manufacturing: A Practical Toolkit, which provides a clear pathway for adoption through its “Scan, Pilot, Scale” framework: identify real operational challenges, test solutions safely, and scale only what proves its value in live environments.

As Kevin told delegates: “Start where risk is manageable, but build deliberately towards industrial impact.” The goal is “repeatable, measurable value”, not more AI for its own sake.

Daniel Isler, Technical Director at Liverpool-based D Squared Product Development, shared how the consultancy has gone from knowing relatively little about the rapidly evolving range of AI tools to applying them within its product development workflow.

D Squared has worked with Made Smarter for around three years, initially using grant support to invest in digital design and manufacturing technologies. More recently, the business turned to Made Smarter's Digital Internship programme to explore AI.

Daniel admitted that at the outset “we didn't really know what was out there”. Digital intern Anoushka Phillips was given the time and space to investigate different AI tools, test them against previous projects and map where they could add value to D Squared's existing processes.

That exploration helped the business identify practical applications across its workflow, from expanding client briefs and supporting research and moodboarding to generating more concepts and transforming the team's own sketches into sophisticated visualisations.

The potential time saving can be significant. Daniel said creating the complex surfaces of a product such as a detergent bottle in CAD could take around a week, whereas AI enabled the team to develop visual concepts from its sketches within hours.

AI can also put those concepts into context much earlier, showing how a product might look on a supermarket shelf, in someone's hand or even in an operating theatre, without first producing and photographing a physical prototype.

But Daniel warned that the speed and sophistication of generative AI can also create a false impression of how far a product has actually progressed.

Using a wearable drug-delivery device as an example, he demonstrated how AI could quickly produce convincing imagery of a product being used in the real world.

The danger, he said, is that “you can very quickly create a facade of a product” that looks finished when much of the engineering work remains to be done.

CAD development, mechanical and electrical design, physical prototyping, testing and refinement remain essential. Indeed, Daniel stressed that these stages are “where all of the learning is”, adding that product developers still need to “come back into the physical world”.

For D Squared, the lesson from experimenting with AI has therefore been as much about understanding its limitations as discovering its potential. As Daniel concluded, these new workflows can “accelerate some areas of the development process such as client communication and visualisation, but aren't currently a replacement for the wide range of design and engineering expertise that is required to deliver a physical product or medical device to market”.

Mark Edwards, Operations Director at Seriun, a North West technology specialist supporting businesses with IT, cybersecurity and AI, focused on the importance of turning enthusiasm for AI into tangible business value.

Working with SMEs to adopt AI within their existing systems and processes, Mark said manufacturers should look for practical opportunities where it can remove repetitive work, speed up processes, improve accuracy or help people make better use of information, and then be clear about what success looks like.

For Mark, the key is being able to demonstrate the return rather than simply adopting AI because the technology is available. That means looking at what a process requires today and measuring what improves once AI is introduced.

As businesses become more confident, the opportunity can extend beyond individual productivity into wider workflows, with AI supporting defined processes and working securely with the data and systems a business already uses.

Mark also challenged manufacturers to think differently about what an efficiency gain actually means.

Saving time should not necessarily be the end of the story. The bigger opportunity is deciding how that additional capacity can be used to create more value.

“What you can do with that spare time, you can usually make the output better,” Mark said. “You can make your customer service better. You can make your product better.”

For Mark, this is an important part of the business case for AI, using the capacity it creates to improve what the business can deliver.

“It's not about replacing the people,” he added. “It's about using the time they get back even more efficiently.”

The Digital Champions Network brings manufacturers together to share what they learn through their digital transformation journeys, exchange practical experiences and build peer-to-peer connections.

Keep your eyes peeled for details of the next Digital Champions Network Forum and how to get involved.