The 3 trends that dominated companies’ AI rollouts in 2025

First, the use of AI for back-end tasks is booming, showing it’s often the boring stuff that can actually move the needle. The second trend isn’t about tech, but rather about people: How companies approach their people is paramount to how AI adoption unfolds. Perhaps the most telling trend, however, is all about initial strategy and motivation. Companies are failing when they lead with AI and finding success when they lead with the problem they’re trying to solve.

Of course, there’s so much more that goes into it—from wrangling data to security and governance. But these aspects of it are shaping AI efforts, for better or worse.

Avoiding AI for AI’s sake

Erik Brown, the AI and emerging tech lead at consulting firm West Monroe, told Coins2Day earlier this year that he’s seen a lot of companies struggle with “AI fatigue” after becoming frustrated with AI proofs of concept that failed to deliver. The common theme among those that fell into this position, he said, is that they explored the wrong use case or misunderstand how AI might (or might not) be relevant for the task. More specifically, they led with the idea that they wanted to pursue AI, rather than with the problem they wanted to solve. 

For one example, he said a client corralled its top data scientists to form a new “innovation group” to figure out how to deploy AI, only to end up wasting tons of resources on ideas that were interesting but didn’t solve any real problems for the company. After his team suggested the firm take a step back and have the business units identify key challenges, the consultants quickly discovered an area where AI could truly help, proved it out by working hand in hand with the business unit, and deployed the solution. 

“I think it’s so easy with any new technology, especially one that’s getting the attention of AI, to just lead with the tech first,” said Brown, echoing an observation I heard over and over again throughout the year, including from company leaders and other consultants helping firms navigate AI transformation. 

One company demonstrating the flip side of this is BigRentz. The construction equipment rental company stayed hyper-focused on the problem it was trying to solve and ended up reinventing its entire business with AI. CEO Scott Cannon told Coins2Day they “didn’t set out to build our company around AI. It just turned out to be the best tool for the job.” What’s more, BigRentz used old-school machine learning only, showing that even in the era of generative AI buzz there’s still value in earlier AI techniques—and why it’s important to find the right solution for the right problem. 

Honeywell is another company that started every pursuit with a clear strategy for what it wanted to accomplish, having created a meticulous framework for guiding its AI development and deployment. It has paid off: Every function and strategic business unit across the company now uses generative AI, and the company has 24 generative AI initiatives in production and 12 more on the way, compared with 16 a year ago. 

“What are the use cases? And can I measure and track them?” CTO Suresh Venkatarayalu told Coins2Day, describing how the company starts with the value add when thinking through any potential AI effort.

Boring delivers results

The idea of avoiding “shiny object syndrome” is solid advice, especially as AI hype quickly jumps from chatbots, to agents, to whatever will come next. Another reason to not chase the latest hype: Many organizations are finding it’s the boring, back-end uses of AI that are truly making a difference. 

Law firm Troutman Pepper Locke is using AI in a wide variety of ways, including creating its own AI chatbot-style assistant for all employees to use. But chief innovation officer William Gaus told Coins2Day the firm is currently finding AI to be most useful for back-end administrative tasks, which he also believes are a great place to start because they’re low risk. 

For example, when the firm was completing its recent merger, his team created an agentic capability to redraft the bios of the incoming 1,600 attorneys, which needed to be updated to include the new firm’s information and match its existing writing style. Gaus described how this made the process drastically more efficient compared to the last time they tackled this task, which took six months of manual work. Overall, the firm saved $200,000 in time spent, he said.

The same thing is playing out in the medical field. Efforts to make a reliable health companion chatbot, for example, have made little material progress. But AI tools are being deployed through the back-ends of the health care system. Doctors are using LLMs to record and transcribe conversations between themselves and patients to generate medical documents, allowing them to engage more with the patient and reduce the burden of time spent on paperwork outside of their work hours. They’re also using LLMs to quickly create synopses of complex medical records and more easily query medical databases.

“Things that we’re taking off of the clinicians’ plate, that are more administrative, I think those are some of the places where we see AI moving really quickly,” Wiljeana Glover, a researcher focused on health care innovation and improvement at Babson College’s Kerry Murphy Healey Center for Health Innovation and Entrepreneurship, told Coins2Day. 

Keeping people front and center

For all the talk about use cases and business strategy, it should not be lost that people are at the center of AI transformation. Whether AI is leading companies to lay off employees is still unclear—and if they aren’t currently, that doesn’t mean this won’t change in the future. Yet AI is already drastically impacting people in their jobs today, from how they’re hired and trained to the tasks and expectations assigned to them. How companies handle the current changes and anxieties about the future has a direct impact on how employees take on AI transformation. 

Perhaps more than any other term, executives I’ve spoken with this year have evoked “change management,” referring to how an organization shifts from its present state to a new, desired form with maximum successful adoption and minimal disruption. 

Honeywell’s other AI lead, SVP and chief digital technology officer Sheila Jordan, warned, “You can’t underestimate it.” Accenture chief AI officer Lan Guan suggested that a business can build all kinds of amazing AI tools that solve business problems, but it’s just as important to make sure your employees are ready and open to using them. Others spoke about needing to bridge the gap between overzealous AI believers (who might be chasing AI for AI’s sake) and AI skeptics. 

A key part of this is company leaders keeping their promises and expectations about what AI can deliver in check. Some developers and software engineers—the first cohort to truly have their jobs turned upside down, thanks to the proliferation of AI coding tools—say they’re frustrated with how many executives are overselling AI and inflating what it can do. Others have felt burdened by unrealistic expectations to produce more code more quickly or to use specific tools, feeling disillusioned by mandates set by executives who don’t understand the day-to-day of their work and are just pushing productivity above all. When these changes are heralded by technical managers with hands-on experience, or even the developers themselves, it often yields a more positive sentiment and better results. 

Even if all the productivity in the world is possible with AI, some executives are wary of what outsourcing too much of the work—especially entry-level work—will mean for the workforce in the near future. For one example, Ryan Anderson, cofounder and CEO of Filevine, a company building AI tools for the legal industry, said he worries about younger lawyers using AI copilots being able to develop their creativity and ability to gather information on their own.

“An overreliance on AI,” he said, “could be just as problematic as the exciting opportunities AI brings.” Finding the right balance should be one key item on the agenda as businesses move forward with AI in 2026.