“This is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning.”
One of Winston Churchill’s many memorable wartime quotes can comfortably be applied to where many businesses stand with artificial-intelligence tools in late 2025 — a mere three years after OpenAI publicly launched ChatGPT. The beginning phase, one of experimentation and “let’s see what this is all about,” is rapidly winding down. It’s time to make AI pay off.
Just how quickly has C-suite sentiment around AI investments shifted? A recent KPMG survey of 1,350 large-company CEOs around the world showed that two-thirds of them expect their AI investments to generate a return on investment in the next one to three years. That’s a remarkable swing from just a year ago, when 63% told KPMG they didn’t expect to achieve an ROI until three to five years.
Here are two examples of how AI is paying off from this fall’s third-quarter reporting season:
- Teams at oil-and-gas company Devon Energy Corp. have been using the technology to analyze the reasons for downtime at the company’s wells from Texas to North Dakota. Chief Technology Officer Trey Lowe told analysts that the next phase of those efforts alone — Devon also is using AI to improve its drilling and gas injection practices — will save the company $10 million next year
- At trucking venture XPO Inc., CEO Mario Harik said shipments per trailer are climbing because teams have started using automated mapping for loading. In addition, AI is adding efficiencies to the local routes trucks are running by optimizing for time, distance and the number of stops. Harik said AI initiatives made XPO 2.5 percentage points more productive during Q3 than it was a year earlier.
The idea that AI is quickly becoming an important part of operations isn’t limited to large corporations, either. A September report from software company Homebase, which helps Main Street-style businesses with scheduling, payroll and other operational tasks, found that 65% of small-firm operators are piloting or using AI tools — and 76% of that cohort describe AI as “very” or “extremely” valuable to their operations.
In short, many teams are figuring out how to get what they want from AI. If you’re not yet at that point — or if you have the nagging thought that your team might be experimenting too much without securing real gains from the technology — experts say a few guiding principles can get you there.
Be true to your core
Most important in an organization’s AI journey is not putting the technology before its North Star, said Charlie Apigian, lead data and AI strategist at LBMC, one of the Southeast’s largest accounting and business consulting firms. Teams should approach AI investments and priorities through a truly strategic lens, but not by thinking AI implementation is the strategy. Instead, he said, the key is to answer the question of why employees love to work at the business.
Productivity boom incoming?
A recent publication from researchers at the Federal Reserve Bank of St. Louis showed that AI tools are being adopted far more quickly than personal computers were in the mid-1980s and the internet roughly 15 years later. It also produced some early evidence suggesting (with a healthy serving of caveats) that areas of the economy adopting AI more widely are seeing greater benefits already: “Industries with higher reported time savings also tended to experience faster measured productivity growth since the release of ChatGPT than pre-pandemic trends would suggest.”
“Get out of the noise and talk about why we are who we are,” Apigian tells leadership teams. “Find that core and then stay true to it. And then adopt AI.”
To find that core, Apigian said it’s critical to involve every discipline in the framing of AI goals and strategies. Leaders from finance, HR, sales and legal should join the boss and the tech champions in those discussions. Doing so, he added, also makes it easier to meet the important goal of getting buy-in from across the organization on Day 1.
Small wins over big swings
Ray Sandza, chief strategy officer at Homebase, said AI isn’t going to run your business for you in 2026. Mind you, he added, that’s not because it can’t yet; it’s more about the typical Homebase client’s customer base not being ready for that. We are collectively still building up our trust in the safety and security of AI systems.
Booking AI successes, Sandza said, is about applying the tech to very specific parts of your organization’s processes.
“It’s not AI for AI’s sake but finding value in every workflow,” he said. “If you do that consistently over multiple workflows, you can string things together.”
For Homebase, that approach means it has started offering some of its more than 150,000 business customers an AI tool that scans timecards for possible errors and omissions. The offering builds on its scheduling and messaging tools and promises to take routine and annoying tasks off the plates of owner/operators.
Apigian framed this phased approach as follows: “Only go as far as it’ll take you today.” That applies directly to the technology but also to how people internally and externally will be able to use it. But it’s important, he said, to generate momentum by launching small-scale projects with a quick pay-off. From there, executive teams can follow up with more expansive ideas also tied to key strategic goals. Otherwise, Apigian said, AI experimentation can meet the fate of fancy dashboards that are used for three weeks and then never consulted again.
Let innovation bubble up
Building on that idea as well as the premise that all parts of an organization should be buying into AI investments from the outset, Apigian and Sandza say it’s important to empower front-line staff — in operations, finance, customer service and elsewhere — to use AI to drive day-to-day productivity.
Sandza said he encourages his teams to be “maximalist” with large language models (LLMs) and see where the limits are. That can lead to dead ends on occasion, but it can also produce real innovation. A case in point at Homebase: A customer support professional used AI to build a new display showing him all the information he really needed when interacting with clients. After putting the tool through its paces, Homebase higher-ups didn’t change a thing about it and rolled it out broadly.
With appropriate security and process guardrails in addition to empowerment, such projects can build broad confidence in AI tools. Apigian said getting that approach right combines the best of both worlds by generating an ROI and encouraging new thinking that strengthens an organization’s culture.
The upshot: Leadership teams looking to get AI “right” need to chart a clear path that emphasizes staying true to strategy and culture while also letting ideas run and prove themselves in a way that may at times be uncomfortable. For Homebase, the ultimate goal is to give back to Main Street owner-operators and employees some of their time. For Apigian and LBMC, it’s to make sure clients don’t lose sight of strategy in their race to adopt new tech.
“I regularly sit across from clients and find myself saying, ‘I don’t have a widget for you,” Apigian said. “‘I want to solve your problem.’”
About the Author
Geert De Lombaerde
Contributor
A native of Belgium, Geert De Lombaerde joined EndeavorB2B in September 2021 to cover public companies, markets, and economic trends primarily for IndustryWeek, FleetOwner, Oil & Gas Journal, T&D World, and Healthcare Innovation. His work focuses on strategy, leadership, capital spending, and mergers and acquisitions, and he also works with Endeavor Business Intelligence on surveys and data projects.
Geert has been in business journalism since the mid-1990s. With a degree in journalism from the University of Missouri, he began his reporting career at the Business Courier in Cincinnati, initially covering retail and the courts before shifting to banking, insurance, and investing. He later was managing editor and editor of the Nashville Business Journal before being named editor of the Nashville Post in 2008. He led a team that helped grow the Post's online traffic by an average of more than 15% annually before joining Endeavor.
Resources
Quiz
Your Competitive Edge, DeliveredMake smart decisions faster with ExecutiveEDGE’s weekly newsletter. It delivers leadership insights, economic trends, and forward-thinking strategies. Gain perspectives from today’s top business minds and stay informed on innovations shaping tomorrow’s business landscape.
