CHOA CONNECTS 2025 Conference Review - Part 3
‘Gravity and Grit’ Part 3: Machine Learning and AI Applications in Heavy Oil
In Part Two of our series revisiting highlights from the CHOA CONNECTS 2025 conference, we reviewed “Operational Efficiency in Canadian Oil & Gas,” with top-of-mind themes including the constant drive for operational excellence, the opportunities and challenges that digital transformation presents in driving this forward, and the critical role of leadership in aligning problems with solutions and managing change with an evolving workforce.
Building on these themes, Part 3 of our series looks back at the conference’s second event, which include three individual presentations delving into the expanding role of artificial intelligence (AI), and how machine learning and other advanced technologies are driving better operating and business decisions. The audience heard from three leading experts: Crystal Lui – Associate Partner Mining & Energy at IBM; Stacey McLennan-Waldal – Director Generative AI & Data Agents at ATB; and Chad Lemke – Executive VP & Chief Commercial Officer at GLJ Ltd. Common threads throughout the discussions included the incredible potential that AI has to drive operational and business decisions, a practical view of AI as a ‘strategic partner’ to embrace in enhancing future workflows, and the ongoing need for human intuition in vetting insights and making business decisions.
Crystal Lui led the group in exploring the possibilities of AI and energy sector alignment, breaking down the role of AI in optimizing the energy value chain and IBM’s work with global oil & gas clients to integrate. She framed the topic of AI in a relatable context, emphasizing industry’s constant drive for improvement and opportunities to leverage vast amounts of data:
“We’re presented with a constant stream of AI information, so it’s more important than ever to separate the hype from what will truly generate value for our business and workforce. Across the energy value chain, our industry has been optimizing for decades. For me, AI is a new set of tools—the next derivative—of that continuous improvement.”
The presentation included a case study of the ‘Exploration Accelerator,’ a web-based tool that assists geologists in searching for and summarizing key insights from documents. These tools are intended to streamline processes by reducing the significant time spent searching for information as opposed to making informed decisions, removing human bias in access and interpretation of data, and addressing expert workforce turnover and the need for knowledge retention.
Stacey McLennan-Waldal introduced the concept of ‘Agentic AI,’ a next step technology which integrates multiple AI agents to perform complex, multi-step workflows to accomplish a goal. Stacey explained the importance of understanding and redefining workflows to leverage the full potential of agentic AI, drawing parallels with the principals of operational excellence, and noting “The factory of the future is powered by AI, but it is built by all of us.” The complex, data-rich nature of the oil & gas business is fertile ground for optimization harnessing agentic AI.Common use cases include predictive maintenance, logistics optimization and environmental monitoring, which can benefit from agents monitoring, analysing and acting in real time.However, the real opportunity may be in parallelling common workflows, using AI to handle background tasks while humans handle decisions from resulting insights. Optimization of steam injection in SAGD is a good example, with complex variables and the potential for AI to continuously analyse, execute workflows and provide insights to reduce decision time and support safer, more reliable operations. In implementing AI, Stacey stressed the need for clearly understanding business objectives, assembling a cross-functional team and planning for data testing. Data science is an experimental and iterative process, and it is critical that leaders get comfortable with failure. Opportunities to integrate AI are scalable, despite a tendency for companies to approach as a large-scale endeavor. Stacey believes AI should be approached as a ‘strategic partner’ to embrace in enhancing future workflows, and that everyone in the industry are data specialists and can be part of the future, stating:
“the future of heavy oil is very intelligent, proactive, and human augmented.”
Chad Lemke wrapped up the session with perspectives on the application of machine learning by industry, presenting a case study on the optimization of multilateral wells in the Clearwater and Manville heavy oil plays. Chad highlighted the rapid growth (over 24 million meters drilled in the Clearwater), significant variability in geology and well performance across the plays, as well as the importance of understanding what factors drive performance to optimize future development. Increasing lateral well length and variability of multilateral well designs are factors adding complexity to evaluations. Machine learning is being used to predict reservoir performance considering variables like oil viscosity, porosity and permeability, and ensuring that pool boundaries and shifting geology are factored into outlooks. ML models currently reconcile 40-50% of performance variance in the Clearwater (higher in some plays), and that with progressive definition of factors and data, correlations are improving. Chad noted the need for ongoing research to improve AI and the handling of geological variability, which could be aided by industry collaboration. Finally, Chad emphasized the importance of human expertise in validating AI insights and for continuous monitoring and adaptation.
Coming up in Part 4…
In Part Four of our five-part series covering the CONNECTS 2025 conference, we explore the event’s third panel, “Heavy Oil Growth Perspectives” in which our industry experts discuss the constraints to realizing Canada’s heavy oil growth potential, as well as opportunities to chart a path forward.
