Industries such as the petrochemical industry, and our broader economy, face significant challenges around productivity, legislation, talent and sustainability. At the same time, AI hype is at its peak, fuelled by a steady stream of headline announcements from the big model providers.
Between those two realities sits the question that matters: where can AI genuinely help?
We had the opportunity to share our insights on AI in maintenance, manufacturing and business optimisation at the European Petrochemical Luncheon in Porto, combining our hands-on delivery experience with academic perspectives on how this can be done rigorously.
Our view is unambiguous: AI will play a critical role in facing these challenges. Predictive approaches to maintenance can shift entire asset strategies from reactive to planned. Data that plants have been collecting for decades, often without using it, holds patterns that translate directly into uptime, safety and margin.
But the industrial context is also where AI hype goes to die. Models that work in slides fail on noisy sensor data, legacy systems and processes where a wrong recommendation has physical consequences. Success here demands domain knowledge first, technology second, and a sober view of what should be automated versus what should support human experts.
That is the conversation worth having with industry: not whether AI matters, but which problems deserve it first.