It seems intuitive: Give AI to the people who are crushing it in your organization and watch them fly even higher. So why does it so often fail to work? Why do so many leaders I talk to complain that their AI transformations simply aren’t delivering the results they had hoped for?
One of the most dramatic and successful AI transformations I’ve seen happened at a bank where the underperforming investment team approached management for help. These folks were struggling and truly needed the help of something like AI. Sure, there was risk involved, but they were willing to take the chance because they had intense motivation to improve.
And when they got that help from AI, they helped create a blueprint for what the technology could do in every other part of the bank. The key was recognizing that AI changes how people work, not simply what tools they use. While a conventional technology rollout can often be taught as a new process or capability, AI requires people to rethink how work itself gets done.
That’s how this bank approached it. They found the people with enough need and motivation to change, then gave them the support to make that change produce a measurable business result. Not just individual productivity gains, but the kind of improvements in revenue, costs and P&L that got the attention of the entire organization.
They started with the scufflers, and it made all the difference.
Why your best performers may be the wrong place to start
One of the biggest paradoxes of the AI age that I’ve observed is that transformation is particularly difficult in successful organizations. Why? Because the people who are best at the current system are often the least motivated to reinvent it. There’s a certain risk-assessment logic at work:
- Employees have agreed-upon goals.
- Changing how they work introduces uncertainty and friction.
- If they believe they can hit their goals without changing, the rational choice is often to avoid the risk.
The result is often adoption theater. People put on a show of participating in the transformation, but for the most part they simply continue to work much as they did before. And AI initiatives stall or fizzle out.
That’s why organizations that have had greater success have started by shifting their efforts to the teams that are under pressure and the business lines that have been struggling, losing ground or discovering that established approaches aren’t working. For these teams, AI isn’t simply a corporate initiative. It can represent a potential way out of a real business problem.
But not every struggling department is going to be fully on board with transformation. Which means need alone isn’t enough. What really helps AI take hold is to identify the rare early adopters and risk-takers on those teams, and get them behind the effort. Then you can really start to see the intersection of AI imperatives take shape:
- A genuine business problem.
- Someone willing to challenge established ways of working.
- A willingness to experiment and take a few risks.
- A situation where success can create measurable business impact.
Identifying those risk-takers is the difference between finding adopters and finding adopters who matter. The ones who can model success not just through happiness or engagement, but through top-line or bottom-line impact.
But finding the right people is just the first step. If the next move is to give them a training course and send them back to work, you’re wasting their potential impact.
Don’t train the scufflers. Work alongside them until they win
Those early adopters are a great resource. But they need concentrated, hands-on support that helps them solve a real business problem to realize their AI potential.
With AI, you’re asking people to work in a fundamentally different way. A conventional training model – with new tools, a two-day training and a handful of coaches or consultants – is not going to get the job done.
AI demands a program with a little more skin in the game. That means support staff working directly alongside the team to apply practices to the company’s actual constraints, solve real problems together and adapt the theory to the specific situation.
For example, when the bank was introducing AI to its investment team, leadership committed to 100 AI experts working alongside 100 client employees, rather than a few consultants attempting to support thousands of employees. This wasn’t coach work. This was actually doing the work together.
The result was a meaningful business win. The transformation changed how the team worked, from producing faster research and underwriting to initiating more client interactions. And the team reversed three consecutive years of market-share losses within 12 months.
That’s the kind of bottom-line result that can be so powerful for an organization. It wasn’t an AI-adoption metric, it wasn’t based on employee satisfaction, and it wasn’t just an impressive number of people trained. It was a tangible business outcome: market share increased.
In other words, it was the kind of measurable win that does more than just prove that transformation is possible. Inside a large company, it can change the behavior of everyone watching.
Turn one team’s win into organizational pull
The investment team’s experience was a perfect example of how to turn AI success into FOMO throughout the organization by sparking curiosity, reducing perceived risk and spreading the capability.
Make the win visible
The retail banking team had previously seen itself as “killing it” and therefore had little reason to change. But seeing the investment team succeed helped change their collective attitude from “Why should we change?” to “How did they do that?” This is the transformation flywheel in action:
- Find a willing risk taker.
- Provide concentrated support.
- Produce a measurable win.
- Make the win visible.
- Let other teams get curious.
- Give those teams experienced practitioners to help them adapt the approach.
It’s an approach to AI transformation that favors pull over push, while simultaneously eliminating the familiar objection that “something like this won’t work here.”
Replicate the people, not just the playbook
But it’s not enough to simply hand the next team a blueprint for AI success and hope they figure it out. Every business unit, after all, has different problems, constraints and organizational dynamics.
Instead, the bank took some of the employees from the investment team and moved them onto other teams. Their lived experience helped them demonstrate the difference between the theory of what a transformation model says should happen and the practice ofhow to adapt that theory to the messy reality.
In doing so, the investment team members became something of a mobile transformation capability, carrying their experience from one successful team to the next. And a relatively small group of early adopters helped influence the entire organization.
Make the CEO the loudest cheerleader
Executive sponsorship in an AI transformation isn’t primarily about speeches, steering committees or generic declarations of support. The CEO’s most valuable role is to provide tangible support and visible recognition.
At the bank, the CEO publicly praised the investment team in a town hall with a straightforward message: after three years of losses, the team was now gaining market share. That recognition turned curiosity into FOMO.
Did the CEO get into the nuts and bolts of the technology? Nope. He simply celebrated what happened, expecting that that would be enough to make the other smart people in his organization want to understand how it happened.
Don’t transform everyone at once
Every organization wants to see big results from AI. That leads to programs that plan the rollout of the technology everywhere at once or in waves, which usually doesn’t work because practical knowledge is spread too thin and the change is challenged by a crowd of detractors.
The more successful AI transformations tend to begin with a minority willing to take the risk. That means identifying people who have a real problem to solve, are willing to change how they work and can produce measurable business impact.
They’re the scufflers. And when they have concentrated hands-on support, a meaningful business result to work toward and proper recognition of their job well done, they can be an organization’s AI vanguard. And then they can move onto the next team, where their example of proof, curiosity and reduced risk will ignite the pull for broader adoption.
They may not be your rock stars now. But if they kick-start your AI transformation efforts, they soon will be.