The productization of human domain expertise is capturing a high-value niche in the industrial software market.
Author: Jagadish Gattu, CEO at UptimeAI
In part 3 of a recent podcast and blog series on “Industrial Systems Engineering in the Age of AI,” ARC Advisory Group Director of Research Colin Masson and industry veteran Rick Bullotta call attention to a new wave of high-value industrial AI applications. These “Domain-specific Innovators,” as they’re referred to in the blog, fall into the ever-widening gap between providers of industry-generic LLMs (the Open AIs of the world) and legacy software companies slapping on a chatbot frontend for the sake of AI marketing.
This middle ground is a continuum with a lot of variability in incremental value to be gained. The solutions that succeed are tackling economically significant problems translating to demonstrable ROI, and they’re doing so by pairing domain expertise with technology innovation to provide vertical or even use case-specific solutions for industrial customers. We strongly agree with their assessment that this is the winning recipe.
Defining Expertise: Depth and Breadth
What we hear consistently from our customers validates the market perception
that there’s a decline in, and a shortage of, expertise in the process industries. They know that they need to capture institutional knowledge before it’s gone, and that digital solutions can help them do this. But what’s inconsistent is how expertise is defined across operating companies, technology providers, and the analyst community.
There’s a common viewpoint that domain expertise = depth (in a particular domain or domains). While this is true, it’s only part true. The way that we view domain expertise, and the way that we capture and codify it, focuses on both depth and breadth. When I say “depth and breadth” I don’t mean that domain experts know everything about every domain. They have a very deep knowledge of one, maybe two, domain(s). But their breadth stems from having working knowledge of multiple domains and understanding the interconnectedness between them.
In my last blog post I broke down the data-driven decision-making process, wherein the rate limiting step is the expert reasoning capability that bridges from insight to recommendation. This step is frequently what stalls delivery of business value because it relies on the few scarce resources that possess both depth and breadth as defined above.
Industrial Examples
In the ARC podcast Rick shares an automotive industry example where a paint defect is linked back to time periods when a totally separate area of the plant was being cleaned sending solvent particulates into the air that made their way to the paint application area. Without the breadth of knowledge of what’s going on in the plant as a whole system, this defect likely would have persisted.
We’ve seen similar eureka moments in our customers. Sometimes a pump issue is a pump issue, but more often than not, when we take a broad system-level view of the problem we find a deeper root cause in the process, adjacent or ancillary equipment. The experience of one of our petroleum refining customers is documented in a case study linking heat exchanger fouling to in-process polymerization and machinery failure.
Conclusions
As ARC Advisory Group points out, the rise of the “domain-specific innovators” technology class underscores the current necessity of scaling expert impact across organizations. This has been and continues to be our “why” at UptimeAI.
By capturing expert reasoning, integrating data across domains, and embedding expertise into operational workflows, organizations can turn the brilliance of an individual into collective intelligence. This shift not only offers huge potential gains in operational efficiency but also positions organizations to thrive in an increasingly complex and digitally sophisticated industrial landscape.
Ready to see Expertise-as-a-Service in action? Get a Demo