--- title: Transcend | Houston 2026 — Session Library description: Watch the full sessions from Transcend Houston 2026 — keynotes, customer use cases, the panel, and the product vision, plus topical clips. --- # Transcend | Houston 2026 Closing the AI Gap with Coordinated Operating Strategies Watch the full sessions from the day — keynotes, customer use cases, the panel, and the product vision — plus short, topical clips you can jump straight to and share. ## Full Presentations All Sessions Keynotes & Vision Customer Use Cases Panels & Discussion Technology Partner & Analyst Customer Use Case ### From 1 to 7: Scaling Industrial AI Across a Refining Fleet **Alex Carlson**HF Sinclair Lessons from scaling Industrial AI across multiple refinery sites — integration, adoption, and organizational change. Watch the Q&A Customer Use Case ### PDH Reactor Optimization Using Deep Learning Process Control **Sam Valleru**INVISTA How INVISTA applied DLPC to improve reactor optimization, stability, and real-time economic performance in a complex PDH operation. Watch the Q&A Imubit Advancements ### Deep Learning Process Control: New Frontiers in Core AI **Prof. Nadav Cohen**Imubit, CTO & President The latest on DLPC — what's new, what's next, and the technical directions pushing the core AI forward. Watch the Q&A Panel Discussion ### Industrial AI and the Coordination Gap **Oron Franco, Chris Harrison, Mahesh Rege** Moderated by Dennis Rohe. Why the biggest barriers to industrial AI impact are coordination challenges — aligning planning, operations, and control — not technology limits. Customer Use Case ### Self Modeling to Improve Daily Decisions at FHR **Brad Touhey**Flint Hills Resources How Flint Hills Resources uses self-modeling to help engineers validate decisions faster and deepen operational understanding. Watch the Q&A Partner Session ### Imubit + Ketjen: Maximizing Refinery Profits with Real-Time Catalyst Intelligence **Henri Tausch**Ketjen A partner's perspective — using real-time catalyst intelligence to compress time to insight and unlock refinery margin. Partner Session ### Partner Session **Neal Biege**FLSmidth Fuller A partner's perspective on enabling service teams to build models directly — putting the platform in the hands of engineers, not just data scientists. Keynote & Fireside ### Industrial AI (R)evolution — Fireside with Gil Cohen **Gil Cohen & Colin Masson**Imubit & ARC Advisory Group Gil's keynote on closing the AI gap with Coordinated Operating Strategies, in conversation with industry analyst Colin Masson on the digital divide and what separates the pacesetters. Imubit Vision ### Product Vision: An MCP-First Architecture **Hieu Bui**Imubit, SVP Product The shift to MCP-first architecture — turning DLPC into an active participant in the agentic workflows industrial customers are beginning to deploy. Watch the Q&A ## Topical Clips **Topical clips are coming soon.** Short, shareable moments from across the day — organized by theme and individually linkable — will live here shortly. All Clips Top Picks Customer Journey Results / ROI Partnership Technology Community 0:21 Top Pick Top Pick ### The front door of a revolution We're at the front door of a new kind of revolution. Even things we talked about two months ago that seemed impossible are now possible. **Tim Yandel**Imubit, CRO 0:08 Top Pick Top Pick ### Training to implementation in two weeks The time from training to actual implementation has been about two weeks. **Alex Carlson**HF Sinclair 0:25 Top Pick Top Pick ### AI as collaborator, not replacement The operator sees this thing is not here to replace me. I turn from babysitting equipment to managing multiple units. **Oron Franco**Prime Six, Operations 0:09 Top Pick Top Pick ### The real friction is cultural AI adoption in industrial environments is no longer optional. The real friction is cultural and operational, not technical. **Ronak Patel**Imubit, Strategic Director 0:06 Customer Journey ### The universal scaling question The biggest question is: how do we adopt these technologies super fast, and especially go from one to many? **Alex Carlson**HF Sinclair 0:09 Customer Journey ### 80% uptime in month one The first month after implementing, uptime was about 80% on the DPC. And then it continued along the same line. **Sam Valleru**INVISTA 0:27 Customer Journey ### From pilot to multi-app rollout We've had one application deployed last year, another online this year, and a third coming within the next couple of weeks. **Mahesh Rege**INVISTA 0:17 Customer Journey ### Change management is challenge #1 Change management is definitely number one. People don't like to be changed. People don't like to be told how to think. **Alex Carlson**HF Sinclair 0:20 Customer Journey ### Incremental wins build momentum Don't go big splash on your first one. You won't get buy-in until they actually see it working. **Alex Carlson**HF Sinclair 0:30 Customer Journey ### Three dimensions of trust I look at trust along three dimensions: credibility, reliability, and outcome-based: are the KPIs moving in the right direction. **Mahesh Rege**INVISTA 0:11 Results / ROI ### Repeatable improvement The DPC and APC combined is slightly higher across all four reactors. That means improvement, and it's repeatable. **Sam Valleru**INVISTA 0:34 Results / ROI ### Stability while pushing constraints When we put the DPC back in, it brought stability and pushed toward the higher end of the constraint. **Sam Valleru**INVISTA 0:36 Results / ROI ### Discovering hidden value The bulk of our data was past the peak of the curve. We'd been running past the ideal location, and there's value in bringing it back. **Brad Touhey**Flint Hills Resources 0:32 Results / ROI ### 100x faster training The time in controller training is on average about 100x faster. If you had 10 hours, it would be like six minutes. **Nadav Cohen**Imubit, Co-founder & CTO 0:08 Partnership ### Why partner vs. build in-house We tried it ourselves. We had an AI guy. Didn't work. We had to find a partner and embed it. **Henri Tausch**Ketjen, CCO 0:38 Partnership ### Platform enables the service team The model builder lets our service team build models. We give them the tools and they run with it; they don't have to be data scientists. **Neal Biege**FLSmidth Fuller, Head of Product & Process Tech 0:21 Partnership ### Compressing time to insight Instead of a month to reformulate, we compress it to two weeks. That's millions a day customers can save by changing earlier. **Henri Tausch**Ketjen, CCO 0:20 Partnership ### A long-term customer relationship Flint Hills has been an early customer since 2018 and has been incredibly helpful in pioneering a lot of our innovation. **Tim Yandel**Imubit, CRO 0:25 Technology ### The foundational DPC patent Model-based RL where your model is an end-to-end neural network, and so is your controller: the foundational patent of DPC. **Nadav Cohen**Imubit, Co-founder & CTO 0:31 Technology ### Addressing core DPC limitations The common critiques of DPC were handling conditions not seen in data and large problems. These directions make fundamental strides there. **Nadav Cohen**Imubit, Co-founder & CTO 0:30 Technology ### MCP as an architectural pillar We're treating MCP as an architectural pillar, not a feature, increasingly exposing DPC models, data, and outputs through it. **Gil Cohen**Imubit, CEO & Co-founder 0:28 Technology ### DPC is deterministic, not an LLM Our DPC models are not LLMs. They're deterministic once trained: same output for the same input, no matter who runs it or when. **Gil Cohen**Imubit, CEO & Co-founder 1:00 Technology ### 2030 vision: one cockpit By 2030, people at all levels will have one cockpit, looking at cross-team execution of operating strategies, down to minutes. **Gil Cohen**Imubit, CEO & Co-founder 0:22 Community ### Community over competition This isn't just about technology. It's about connecting the dots and ideas that better serve your jobs, woven into how you work. **Tim Yandel**Imubit, CRO 0:23 Community ### The digital divide About 12-13% of companies are leaders deploying at scale. If you're not a pacesetter, yes, you are falling behind. **Colin Masson**ARC Advisory Group, Analyst 0:17 Community ### The value of peer learning What's special is connecting on the ideas and challenges we're all facing across different verticals. **Tim Yandel**Imubit, CRO Link copied