TNChE Asia 2025 took place last week 20-22 May in Pattaya, Thailand. The event gathered chemical engineers, energy leaders, and researchers focused on advancing the performance and resilience of petrochemical and energy-intensive operations. Topics centered around process control, plant modernization, and how to build more responsive systems in volatile conditions.
The Beyond Limits team in Asia joined the conference to contribute to the discussions around operational support, machine reasoning, and practical industrial AI. Our focus was clear: help operators, engineers, and planners make better decisions using systems that reflect both data and domain knowledge.
We weren’t there to pitch theory. We shared direct outcomes from our work with refineries and chemical plants. These stories are more than case studies. They reflect the real challenges companies face—and the tools they’re starting to trust.
Petrochemical facilities operate under continuous pressure, maximize throughput, minimize risk, and adapt to changing conditions without missing targets. Every shift presents new data, new variables, and new risks. At TNChE Asia, this reality was front and center. Most teams in attendance acknowledged the same challenge: they’re collecting more data than ever but struggling to turn it into actionable insight.
At the BP Whiting Refinery, Beyond Limits addressed that head-on. Together, we launched a cloud-based cognitive AI advisor that helps operators manage complexity in real time. Instead of waiting for lab results or reacting to alerts, operators get recommended actions based on current process data and expert knowledge encoded into the system.
In its first year, this solution drove over $10 million in added value—primarily from higher product yield and improved plant utilization. It also helped align operators around shared goals, reinforcing standard operating procedures and reducing plan-vs-actual delivery gaps by 17%.
The tool wasn’t designed to replace expertise. It captured it, structured it, and made it available to teams throughout the refinery. Now in its next phase, the solution is evolving to support other personas—including planners and engineers—through predictive modeling, hybrid AI applications, and virtual sensors for real-time insight.
What stood out at TNChE Asia 2025 wasn’t the push for futuristic technology. It was the demand for grounded tools that help people do their jobs better—tools that fit into the plant, not just the pilot. Attendees wanted to know: How do we help operators respond faster? How do we give engineers better visibility? How do we reduce complexity instead of adding to it?
That’s exactly where our Operations Advisor fits.
It doesn’t sit on the sidelines or requires new infrastructure. It becomes part of daily operations. Frontline workers use it to make decisions with more confidence. Engineers use it to understand why gaps exist between plan and performance. Planners use it to adjust models in real time.
It’s not an overlay. It’s a new way to structure shared understanding, linking data, domain knowledge, and recommended actions in one place. And it’s built to scale: logic-driven, operator-facing, and ready to be audited at any time.
This approach isn’t theoretical. It’s already in place and is expanding to support hybrid AI use cases across planning, modeling, process optimization and energy management. What we saw at TNChE Asia 2025 reinforced this is exactly the kind of solution the sector needs now.
TNChE Asia 2025 didn’t just spotlight new ideas. It gave us a clear signal on where the industry is focused—tightening margins, managing emissions, and gaining better control over increasingly complex operations.
At Beyond Limits, we don’t talk about the future of AI in industrial environments—we build it. With engineers. With plant managers. With the people doing the work every day.
If you're leading operations in petrochemicals, energy, or large-scale chemical production and you’re looking for support that’s grounded in execution, we’re ready to talk.
The challenges are real. The solutions can be, too.
