Tim McLain, Co-founder of Lexin Solutions
New research from James Hurman and Klaviyo finds heavy discounters grow the slowest of all, right as retailers plan their EOFY sales
Artificial intelligence has rapidly become one of the defining business conversations of the decade. Every enterprise software company now has an AI strategy. Every platform is introducing intelligent features, and every board is asking how AI can improve productivity, reduce costs and create competitive advantage. The assumption has been that as AI models become more capable, enterprise adoption will naturally accelerate.
That may be true for low-risk productivity tools, but industrial organisations operate in a fundamentally different environment. In mining, manufacturing, utilities and heavy industry, AI is no longer being asked to summarise documents or automate routine administration. It is increasingly influencing decisions that determine whether millions of dollars remain tied up in inventory, whether critical assets stay online, and whether maintenance teams have the right materials available when equipment fails.
The question, is no longer whether AI can generate an answer. It is whether businesses can trust that answer when the commercial consequences are significant.
This is where I believe much of the enterprise AI conversation has lost its way. Too much attention has been placed on the sophistication of the technology itself and not enough on the expertise that sits behind it. Intelligence is becoming increasingly accessible. Domain expertise is not.
Having spent more than two decades working across global organisations including BHP, Holcim, Incitec Pivot and OneSteel, I have seen first-hand how complex industrial operations really are. Maintenance, repair and operations (MRO) supply chains rarely fail because organisations lack technology. They fail because operational knowledge is fragmented, business processes have evolved over decades, and critical decisions are often being made using inconsistent or incomplete data. That experience ultimately became the foundation for Lexin Solutions and MiLi. We didn’t begin by asking what AI could do. We began by asking what problems the industry had been trying to solve for years that technology still hadn’t addressed.
As industrial organisations accelerate investment in AI, those challenges are becoming more visible. Gartner estimates poor data quality costs organisations an average of US$12.9 million every year while McKinsey reports that 80% of organisations cite data limitations as one of the biggest barriers to scaling AI. Those figures reinforce an uncomfortable reality: AI cannot create certainty from unreliable information. It simply processes poor-quality data faster and produces recommendations that appear increasingly confident, regardless of whether they are correct.
Across asset-intensive industries, those risks are amplified. Duplicate material records, inconsistent naming conventions, obsolete supplier information and fragmented maintenance histories exist in almost every large industrial organisation. Applying AI to those environments without first understanding the operational context does not eliminate complexity. In many cases, it accelerates it. That is precisely why Lexin was built differently.
MiLi was designed by people who have spent their careers inside warehouses, maintenance teams and supply chain functions, not by observing those environments from the outside. Every workflow, recommendation engine and optimisation capability reflects decades of practical experience managing indirect materials across some of the world’s largest industrial businesses. Rather than asking customers to adapt to generic software, the platform captures operational knowledge that already exists inside organisations and applies it consistently at scale.
An experienced maintenance planner understands why a critical spare part cannot be optimised using the same logic as finished goods inventory. A procurement specialist understands why supplier lead time may be more important than purchase price. An engineer understands that selecting the wrong component can compromise reliability, safety and production. These are commercial judgements built over decades of operational experience, and they are exactly the types of decisions enterprise AI must support rather than replace.
The businesses creating the greatest value are recognising this shift. Rather than removing experienced people from decision-making, they are embedding their knowledge into enterprise software so it becomes repeatable, scalable and consistently applied across thousands of operational decisions every day. AI provides the speed and analytical capability. Domain expertise provides the judgement.
Increasingly, enterprise buyers are recognising the same distinction. Organisations are asking different questions of technology vendors. They want to know who built the software, what operational expertise informed the product, how recommendations are governed and where human accountability remains embedded within the decision-making process. Trust, explainability and industry credibility are becoming just as important as model capability because enterprise AI is no longer judged by how intelligent it appears, but by whether organisations are prepared to act on its recommendations. This represents a significant shift for enterprise software.
The winners in the next generation of AI will not simply be the businesses with the largest engineering teams or access to the latest foundation models. They will be the organisations that combine artificial intelligence with decades of operational expertise, trusted data and disciplined governance. In enterprise environments, particularly those supporting critical infrastructure, technology alone has never been enough.
Artificial intelligence will undoubtedly transform industrial operations over the coming decade. However, the businesses that create the greatest commercial value will be those built by people who understand the industries they serve because, ultimately, the most powerful AI is not the one that knows the most. It is the one that understands the problem well enough to make better decisions.
