Over the last few years, almost every industry has been experimenting with AI. The convenience and fuels business is no different. But the conversation is starting to change.
We are moving past the stage of asking whether AI can do something impressive in a demo environment. The bigger question now is whether companies can actually make it work inside real operations across stores, supply chains, trading desks, dispatch groups, and field teams.
That was the focus of a recent episode of the Future of Convenience podcast, “AI Gets Real: Pilot Projects to Performance at Scale,” featuring Doug Haugh, Chairman of NewTide AI, alongside Jeff Rubin, Senior Vice President of Fuels at Upside.
The discussion covered where AI adoption actually stands today in convenience retailing and fuels distribution, and why the next few years are going to look very different from the last few.
The Industry Has Been Careful. And That Might Be a Good Thing.
One point Haugh made early in the conversation was that our industry has traditionally been slower to adopt new technology compared to others.
“Absolutely, we’re a technology laggard in many ways,” Haugh said. “Because we can be. We provide an irreplaceable product and service, so we’ve had the luxury of time to filter what’s real from what’s hype.”
There’s actually something important in that mindset.
A lot of industries rushed headfirst into AI pilots over the last several years. Some produced value. A lot didn’t. Many companies discovered that building a flashy pilot is one thing. Getting it to work reliably inside a real business is something completely different.
That’s especially true in fuels and convenience retailing, where operations are complicated, margins matter, and reliability is everything.
Rubin talked about how much faster startups are able to adapt compared to larger organizations.
“It’s going to take them time to adapt to today’s new world,” Rubin said when describing larger company environments.
And that’s really the challenge the industry is wrestling with right now. How do you move quickly enough to stay competitive without creating chaos inside the business?
AI Pilots Are Easy. Scaling Them Is Hard.
One of the biggest themes throughout the conversation was that most companies are now realizing the hard part of AI isn’t the model.
It’s making it work in real operations.
The reality is that most organizations already have messy data, disconnected systems, manual workflows, and teams overloaded with day-to-day execution challenges. AI exposes those problems very quickly.
That’s one reason so many pilots struggle to scale.
As Dan Munford recently said:
“Most AI pilots don’t fail because of the model. They fail because organizations aren’t ready to scale.”
Doug described NewTide’s approach as being built around three connected layers:
- AI agents that help users interact with the system
- A contextual enterprise data layer called Data Helm
- A workflow orchestration platform called Shipyard
The important point is that AI only becomes useful when it’s connected to real operational context.
Most companies already have valuable information spread across ERP systems, contracts, shared drives, spreadsheets, emails, and operational systems that don’t naturally talk to each other. At NewTide AI, that’s exactly the problem our Data Helm platform is designed to solve by pulling together both structured and unstructured enterprise data, understanding how those systems and data points relate to one another, and exposing that context to AI agents and workflows in a way that is actually usable inside day-to-day operations.
Start With Something Useful
Another part of the discussion that resonated was Haugh’s recommendation for companies that are still early in their AI journey.
Don’t start by trying to reinvent the entire business overnight. Start with something practical. Something useful. Something that solves a real operational problem.
“Pick a boring but useful use case to start with,” Haugh said. “Learn how this stuff works, how to get value from it, and how to build real automation inside your business.”
That advice reflects a reality many companies are now discovering firsthand. AI adoption is not just about buying a model or standing up a chatbot. The hard part is figuring out how the technology fits into existing workflows, how it interacts with operational systems, and how employees actually use it day to day.
A lot of organizations made the mistake early on of treating AI like a massive transformation initiative from day one. In many cases, that created large budgets, unclear expectations, and pilot programs that never made it into real operations.
The companies seeing the best results are usually taking a much more disciplined approach. They are starting with targeted use cases that create immediate operational value. Things like automating repetitive workflows, improving visibility across disconnected systems, helping employees access information faster, or reducing operational noise for frontline teams.
Once companies see measurable value, the conversation changes.
Instead of asking teams to fund another experimental technology initiative, the business starts reinvesting in something that is already improving execution and producing results.
That approach has been central to how NewTide AI has approached deployments across supply, trading, distribution, and retail operations. As Haugh explained during the podcast, 2025 was focused on proving the platform across every major link of the supply chain, from global trading organizations to midstream logistics operators to convenience store retailers themselves.
The goal was not to create isolated demonstrations. It was to validate that AI could operate inside real-world environments where execution, reliability, and operational consistency matter every day.
That distinction is becoming increasingly important as the industry moves beyond experimentation and toward making AI part of normal business operations.
The Real Opportunity Is Operational Execution
The conversation eventually shifted away from technology and toward what may actually matter most: execution.
Haugh made the point that after spending years operating in the industry, running stores, working through logistics and supply chain challenges, and managing operational complexity firsthand, you realize quickly that technology only matters if it actually helps operators execute better.
“A lot of technology vendors are fascinated with the shiny new thing,” Haugh said. “But if operators can’t see how to fit it into their business while juggling everything else they have to do, nothing happens.”
That’s where the conversation around AI is maturing.
The most valuable AI systems are probably not going to be the loudest or flashiest. They’re going to be the ones that quietly eliminate friction, reduce distractions, improve consistency, and help teams focus on customers instead of constant operational noise.
Haugh described that future as “quieter” operations.
Fewer emergencies. Fewer stockouts. Less chaos. Stores that simply run better.
And for an industry dealing with ongoing labor shortages and execution challenges, that matters.
AI as a Digital Workforce
The conversation also touched on a broader shift in how companies are beginning to think about AI not simply as automation software, but as a “digital workforce” that augments human teams.
“I’m not using this to simply automate a process,” Haugh explained. “I’m using it to create a digital workforce that complements my human workforce in a very powerful way.”
That’s an important distinction.
The goal isn’t replacing people. The goal is helping good operators perform at a higher level by reducing repetitive work, improving visibility, and allowing teams to focus where human interaction matters most.
Rubin pointed to examples already emerging at the store level, where AI and automation are helping improve customer engagement, coaching, loyalty execution, and operational consistency directly through POS-integrated systems.
What Happens Next
The convenience and fuels industry is still early in this transition. But it’s becoming increasingly clear that the companies creating real value from AI will not necessarily be the ones running the most pilots.
They’ll be the ones that:
- integrate AI into real workflows,
- connect it to operational data,
- focus on measurable execution improvements,
- and scale carefully but deliberately.
As Dan Munford summarized during the discussion:
“We can’t see the whole future, but we are starting to see parts of the future emerge.”
And increasingly, that future looks a lot less like experimentation and a lot more like real operational execution at scale.
NewTide builds an enterprise AI platform for convenience, fuels, and other operations-heavy industries. This conversation aired on the Future of Convenience podcast, recorded at NRF 2026. To learn more about how NewTide helps organizations move from AI pilots to production, visit NewTide.ai.
Source: Carolyn Schnare, “AI gets real: pilot projects to performance at scale with Doug Haugh and Jeff Rubin,” Global Convenience (Global Convenience Store Focus), 17 March 2026. https://www.globalconvenience.com/features/ai-gets-real-pilot-projects-to-performance-at-scale-with-doug-haugh-and-jeff-rubin/
Additional commentary adapted from Dan Munford and Doug Haugh via LinkedIn.