What the Firms Getting Real Value From AI Do Differently
The appeal of AI implementation to an owner or managing partner is easy to understand. It means freeing up talent to focus on high-value, high-judgment decision-making and potentially increasing the capacity of the firm without a corresponding increase in overhead. It means finally unlocking the full value of the technology the firm is already paying for, while also creating the space to rethink the value proposition the firm presents to its clients. The question becomes one of how to get implementation right, because it’s clear that the magnitude of the potential is matched by the speed of change, and everyone intuitively feels that competitive advantage will go to those who get things right.
At the same time, in every firm that announces it’s adopting AI, four different reactions are happening. One person is relieved that a tedious task is finally going away. A second has never once stopped to think about their own job as a set of steps that could be looked at individually. A third has already sat through a technology upgrade that promised to change everything and didn’t. And a fourth is quietly certain this is the first step toward not needing them at all. That range of emotions, in one meeting and multiplied across an organization, captures the real challenge facing any successful AI implementation.
The Wrong Kind of Change
A recent global survey of CEOs found that roughly half report seeing no real financial return from their AI investments so far. Does that mean the transformative potential of AI is overstated? Or does it just come down to how it’s being deployed?
Most firms plan for the wrong kind of change. When a transition like this gets managed at all, it usually gets managed at the level of tasks: which processes will change, what new tools people need to learn, how long training will take. That’s real work, it matters immensely, and it is often not done properly. But it is also not the change that often derails success.
Research published in MIT Sloan Management Review describes a second and much quieter kind of disruption: role-level change, the way a transition can unsettle how someone understands their own value inside a firm. Task change asks whether you can learn something new. Role change asks whether you’ll still matter once you have. Most planning stops at the first question because it’s the
one people expect and can prepare for. It’s the second question that strikes to the heart of an individual’s value in an organization, and it’s the one most people don’t realize they need to tackle.
The people being asked to adopt AI are often, in their own areas, experts. Many have spent years building the judgment that makes them good at what they do. Being protective of that isn’t a failure to embrace change. It’s the reasonable reaction of someone who worked hard for what they know.
No previous technology has arrived carrying this much prior debate, both hopeful and anxious, so people show up already convinced of what it means before they’ve had a chance to see how it actually changes their own work. That context doesn’t create the role-level disruption described above, but it does mean that disruption now lands at a whole other magnitude.
This Has Been Done Before
None of this is new, and none of it is unmanageable. What separates the firms that get through a transition like this from the ones that stall out six months in is whether they do the real diagnostic and process analysis work up front, and whether they keep the human dimension of the change at the center of it throughout. It’s whether leadership understands the change management challenge before them and takes the steps necessary to address it.
The instinct in most rollouts is to meet skepticism with enthusiasm and meet fear with reassurance. Both responses skip past the point. Skepticism from someone who has already sat through an overpromised tech platform is not an attitude problem. It’s information about what will and won’t work this time. The same is true of fear: a person who has watched their expertise treated as replaceable before has good reason to ask hard questions now. A firm that treats these reactions as data worth designing around, rather than resistance to manage past, gives itself a real shot at adoption that actually sticks.
What Taking It Seriously Looks Like
What that looks like in practice is not complicated, even if it’s rarely done. Start by mapping how a job actually works: which parts are repetition, which parts are judgment, and which parts nobody has ever separated out before because no one ever had reason to do so. Say the fear out loud in the room instead of covering it with forced enthusiasm. Treat the person who’s skeptical because of the last overpromised tech platform as someone with useful information, not a mindset to correct. Ask the people doing the work what they think the technology should and shouldn’t touch, before deciding for them.
What it requires is doing the right work at the very start of any technology deployment, and doing it with people who take the cultural and change-management side of the transition as seriously as the technical side.
The firms that keep asking these questions as things keep changing are the ones that end up with real ROI. The ones that treat it as a one-time announcement are the ones still asking, a year from now, why the investment never paid off.
Sources: Hal Gregersen and Roger Lehman, “How Organizational Change Disrupts Our Sense of Self,” MIT Sloan Management Review, Summer 2021. PwC, 29th Annual Global CEO Survey, January 2026.
