Industrial AI - real value today for complex engineering challenges

The conversation around Artificial Intelligence (AI) is louder than ever, fuelled by rapid advancements and massive investments. Beyond the hype of large language models (LLMs), what does 'true value' in AI actually look like? Dominic Vergine, CEO of Monumo, discusses the current state of the AI market, where venture capital should focus, and how Monumo is using precise, Industrial AI to deliver immediate, measurable value in the complex world of industrial engineering.

As a founder and CEO in the AI space, what is your take on the current AI bubble narrative, and where should VCs be investing?

The question of whether it's a bubble remains to be seen, but in any situation where a truly radical and exciting new technology is just beginning, there will be winners and losers.

The companies that will succeed are the ones with their own technology. Specifically, for smaller companies, success lies in being genuinely focused on using AI to deliver true value to customers in whichever sector you operate.

The ones that won't succeed are those that have simply built their product or service on top of larger existing LLMs like Claude or ChatGPT. My view is that these companies are potentially vulnerable. The foundational layer will eventually "eat the value" and probably offer the service itself. You must have your own data, your own core technology, and a real focus on a real need, not an imagined one. I believe that distinction will become crystal clear in the next three to five years.

What does true value mean in a practical sense when integrating AI, especially in a sector like engineering?

For Monumo, true value is about real cost saving. We now have examples of how we’re saving customers a significant amount of cash per motor system. When those customers are selling millions of units, there is significant potential for cost reduction. That’s the absolute priority.

The speed or efficiency AI offers is secondary to that - almost icing on the cake. If we were just offering speed, the barrier to disruption would be challenging. Efficiency is no doubt a benefit, but it's not enough to justify the required internal change. I suspect that is why, across the board, enterprise buyers of AI are struggling to recognise ROI at the more 'clerical' end of AI - faster emails, better-managed CRM systems, etc.

The other essential component is accuracy. A lot of AI is criticised for getting 90-95% of the way there. But whether it’s a legal document or a complex engineering design, 95% isn’t good enough. You can’t spend all your time going through it with a fine-tooth comb, looking for the 5-10% that is wrong or hallucinated. You need to prove the accuracy of the AI’s output to build the necessary trust.

Monumo started with a focus on the EV space. Why the shift to a broader engineering market focus now, and what have you learned?

We initially focused on electric motor systems for the EV market to prove product-market fit. We did that relatively quickly, and we are already working with major global OEMs to deliver real value.

As is often the case, the demand for what we offer has pulled us into other markets. Our list of companies for pilots and commercial projects right now isn’t just automotive. We are in active conversations with companies in the domestic appliance, HVAC (heating, ventilation, and air conditioning), heavy machinery, truck, and drone markets. If we just focused on the electric motor, we would be playing in a market worth tens of billions of dollars, but we believe there is significant potential beyond it.

The logical next step is also to expand the systems we model. We’re already moving beyond just the motor to incorporate the mechanical gearing, inverter & control software, and further into areas like battery architecture and hybrid propulsion. This will take us deeper into other physical domains such as aerodynamics and require additional modelling approaches such as CFD (Computational Fluid Dynamics). This allows us to go one step further in system-level optimisation, delivering value for our customers’ next production design, not something that might appear in R&D in 10 years.

How are AI and Machine Learning being received by these traditionally engineering-heavy businesses? What are the biggest challenges to adoption?

We had a very senior engineer from a major global automotive company reach out to us, and when we showed him what we could do, he simply said, "This just seems like magic. I’m not sure I can believe it."

That’s hugely exciting because we know we can prove it, but everyone will go through that period of scepticism. The difference we deliver is so radical that this reaction is understandable. This is what happens when you have a truly disruptive shift in an industry. Scepticism is also vital in high-risk, regulated markets as it forces you to prove the accuracy.

The reality is that what we are doing isn't magic; it's about optimising how all parts of an engineering system work together, but it is very new and exciting. We’re delivering optimal engineering systems AND significant cost reductions. Our customers can readily check our designs with their usual tools, giving them the same high level of confidence in our AI-generated designs as they would in their own designs.

What is the biggest takeaway about Monumo and the future of industrial AI?

The biggest focus is on delivering real value today, not tomorrow. We have a strong, foundation for building AI for engineering, it’s not 95% of the way there; for the electric motor systems we are focussing on today it’s 100% of the way there. If you try to build a successful business on a fundamental foundation that only gets 95% of the way, it’s going to get worse, not better. Industrial, physical AI - whether for engineering, medical, or other high-stakes fields - must deliver complete accuracy and be ready to use. That's the difference between clerical AI, which takes up most of the press column inches at the moment, and the profound change that industrial AI is bringing to the world.