Over the past several months, my team and I, alongside AI and technology experts, have engaged in candid conversations with energy executives across the downstream value chain. These roundtables weren’t about showcasing innovation for innovation’s sake. They were about listening. What emerged was a grounded, pragmatic view of what it really takes to make AI work in the messy, margin-tight realities of refined fuels.
The executives didn’t come to talk about roadmaps or theoretical futures. They came to talk about results. They want to know how AI can improve pricing accuracy, sharpen demand forecasts, and simplify complex decisions across marketing and supply.
“I’m interested in how AI helps us make faster, better decisions, not just automate what we already do,” one marketing executive told us.
That sentiment reflects a broader shift in the sector: from exploration to execution. Deloitte’s 2025 Oil and Gas Industry Outlook confirms this trend. Companies are favoring short-cycle, outcome-driven AI projects over sweeping digital transformations. It’s less about revolution, more about traction.
This nimble approach is well suited to a volatile market. Pricing and trading leaders emphasized the need for AI tools that anticipate demand swings and margin pressure before they hit, not tools that simply explain what already happened.
As one of them joked, “I don’t need AI to tell me what already happened.”
The proof is in the profit. Based on customer interviews, historical analysis, and pilot studies, DTN found that oil and gas companies using AI in pricing and margin management can achieve combined sales and margin increases of 1 to 2%. In a business where a single basis point can win or lose a bid, this is meaningful.
One of the most consistent themes across the roundtables was the value of precision through iteration. Leaders aren’t chasing perfection. They’re chasing repeatable wins through small, focused improvements that compound over time.
Projects that delivered early value weren’t massive overhauls. They were tightly scoped efforts like automating daily pricing updates, improving demand forecasts for key terminals, or reconciling bill-of-lading data. These initiatives didn’t transform the business overnight, but they built credibility and momentum.
As one executive put it, “Don’t chase perfection. Chase repeatable wins.”
Organizations that defined ROI metrics upfront, including margin capture, reduced rework, and faster decision cycles, were the ones that sustained progress. The lesson: progress is iterative, and iteration is the path to precision.
AI is beginning to reshape how cross-functional teams collaborate in downstream operations. Supply, trading, pricing, sales, and scheduling teams are starting to share real-time data through AI-driven platforms. Machine learning models pull from refining units, logistics systems, and market feeds to create a shared analytical layer.
But the challenge isn’t building dashboards. As one executive quipped, “Dashboards are so 2000.” The real challenge is agreeing on what the data means.
McKinsey research shows that integrated data platforms can cut decision times by 30%. But that efficiency depends on alignment, not algorithms. Without trust across teams, AI risks amplifying confusion instead of clarity.
“If you don’t fix the silos,” one executive warned, “AI just makes the silos faster.”
If there was one universal truth from every roundtable, it was data is still the main barrier. Nearly every executive mentioned the challenge of data hygiene, including mismatched product codes, inconsistent terminal data, and siloed systems, that limit model accuracy.
“The models aren’t the issue, our data is,” said one analytics manager.
Another added, “Getting everyone to agree on the data feeding it is the hardest part.”
The goal isn’t twelve perfect answers. It’s one imperfect answer that everyone can act on as the single source of truth.
Many organizations are still in what one participant called the “structured learning phase.” They are testing models, cleaning data, and proving value. Some have built AI scorecards to quantify impact through faster pricing cycles, reduced manual work, improved margin capture.
“If you don’t show value early, enthusiasm fades fast,” noted one executive.
This grounded mindset is helping separate real progress from theoretical ambition. It’s not about chasing transformation. It’s about accelerating learning to enable progress, one iteration at a time.
After months of conversations, one theme rose above the rest. The companies gaining ground with AI aren’t chasing innovation. They’re mastering operational decisioning.
Speed and accuracy now define downstream competitiveness. The ability to make the right call faster than the next marketer or trader is becoming a new measure of strength.
Yes, the hurdles are real, from data chaos to legacy systems and organizational inertia, yet optimism remains. The roundtable conversations revealed that downstream oil and gas companies are moving closer to a future where AI is embedded in the decisions that move barrels, margins, and markets.
The next chapter of AI in energy won’t be written by algorithms. It will be written by leaders who see precision not as a destination but as a discipline. And the companies that win won’t be those chasing perfection, but those accelerating learning through repeatable wins.
Ken Evans General Manager, Energy and Refined Fuels, Ken Evans leads the energy and refined fuels business assuring our solutions continue to deliver exceptional value to customers in this sector. He has more than 30 years across the oil and gas industry, including Global Vice President for Oil & Gas at SAP prior to joining DTN. He has been actively engaged in various youth leadership development and leadership positions in his community and church. He earned his bachelor’s degree in chemical engineering from the University of Colorado.
