When leadership calls for an AI pivot, two product leaders show what it looks like to lead through it — one from the receiving end, one from the driver's seat.
Your CEO just walked out of a company offsite with a new top priority: AI. Everything. Now. 90 days.
Your roadmap (the one grounded in months of discovery and customer interviews) is gone. Your team is asking whether they're ignoring real problems for hype. And you're the one who has to make sense of it.
That's the scenario at the center of Episode 5 of the Product Leaders Lab. I talked to two product leaders who've lived this exact moment, from completely opposite sides. Lauren Schuman, VP of Product Growth at Close, had the pivot handed to her four weeks into a new job. Shannon Vettes, CEO and CPO at Usersnap, was the one driving the pivot, against a team that wasn't ready. Same pressure, different starting points, genuinely different approaches. Here's what they learned.
Most AI pivot conversations focus on the what: which features to build, which models to use, what the roadmap looks like after the dust settles.
But the harder problem is the how. How do you lead a team through a shift they didn't ask for? How do you move fast when your team is confused, and customers haven't asked for the thing you're suddenly building? And how do you stay grounded in real customer problems when everything above you is pointing in a different direction?
This is a leadership problem, not a product problem. And the teams that get it right aren't the ones with the best AI strategy. They're the ones whose leaders create clarity before they have certainty.
Lauren walked into Close expecting to lead a PLG strategy. Growth systems. Activation. That was the role she signed up for.
Four weeks in, she was sitting in a company offsite when her CEO announced the company was going all in on building an AI voice agent. Not as a side experiment, she told me. "This was an identity-level shift for the entire organization."
The team she'd just inherited had finished building their roadmaps. Suddenly, those plans didn't exist anymore. And Lauren, with four weeks of context, was the one who had to help everyone make sense of it.
"There were a lot of teams asking: what about this? What about that? Customers aren't asking for this, but they're asking for that. We know this solves real problems. Are we ignoring these tangible asks in favor of something that people aren't asking for?"
She was honest that she felt both sides of that tension herself: excited about the opportunity to work on AI-native features, but not anticipating the "all in." Both emotions, at the same time.
She didn't wait for clarity from above. One of Lauren's most important early moves was deciding to stop looking upward for answers.
"Stop waiting for clarity. Go ask the questions. Have a point of view and create clarity, because you're asking people to create clarity and they may not even know. They have a concept in mind, but they don't know the details. So it's your job to make that clarity happen."
She also made a counterintuitive call when it came to staffing: she didn't bring in an outside AI expert to lead the charge.
"Nobody's done this yet. Yes, there are people who have experimented, but nobody has really nailed it. So the advantage from bringing someone in from outside? It's not as clear as you'd think. We had people who deeply understood our product and our customers. That combination matters more."
She stayed with the existing team, ran a dual track (supporting current work while quietly building the future roadmap), and got her hands into the product directly. She built her own voice agent. She broke things. She learned what the technology actually felt like.
Then she did something she admitted she should have done sooner: she went directly to customers.
"I waited longer than I should have to just be in front of customers. Originally I was consuming the research my PM was doing. Then I finally said: I'm canceling a bunch of my other meetings and I'm going to every alpha customer interview. In the last two weeks, I've been part of at least 12 conversations. And there is nothing that replaces being there."
Twelve customer conversations in two weeks. That's not delegation. That's a decision to anchor the pivot in reality, fast.
Lauren's path asks you to lead through fog. You don't need to see the destination clearly. You need to keep moving and keep asking.
Shannon's situation looks like the mirror image of Lauren's.
Lauren had a mandate handed down. Shannon had a conviction, and had to fight for it. She was the one at Usersnap trying to get the team to move toward AI. And her team pushed back.
The origin of Shannon's conviction is almost mundane: she was annoyed. Genuinely, personally frustrated that she was still manually tagging customer feedback in her own product.
"I'm not someone who is easily annoyed. You have to try hard. So that's maybe why I use my own gut check for a lot of our strategic directions. If I'm annoyed by it, man, it must be really bad."
She went to her team and asked why they weren't automating this with AI. Their response: they had qualms. Ecological concerns, security and compliance questions, and a more fundamental challenge: did this actually need AI, or was there a better way?
Here's where Shannon made a move that I think most leaders get wrong. She didn't override the resistance. She used it.
"Their ability to disagree with me is my biggest asset in decision-making. And I will never tell them to stop disagreeing with me. I will always encourage their critical thinking. Because it is what fine-tunes the product decision."
Those "qualms" weren't just friction. They shaped the architecture. Because of the team's ecological and security concerns, Usersnap moved away from a query-based approach and toward a more conservative, batch-based AI system, which turned out to be better for compliance too.
Her team's resistance didn't slow the pivot. It made the product better.
Shannon also found a way to accelerate without hiring blind: she tapped into the central AI team at SAS Group.
"They were able to share what they already knew, what they'd done, what worked, what didn't. We leveraged a lot of their trial and error. I feel like we had a real advantage, an unfair advantage as a small business."
When she did decide to move forward, she put her name on it publicly.
"I kind of pushed forward and decided to do it anyway. Probably the biggest decision I made as a leader in the first six months was: okay, everybody get on board. We're doing this, sorry. And if it fails, it fails and I'll take responsibility for its failure."
That kind of ownership changes the dynamic. It shifts the question from "are we doing this?" to "how do we make this work?" It creates the psychological safety for the team to move.
Shannon's path asks you to lead from conviction. You don't need permission. You need a clear point of view and the willingness to own what happens next.
Lauren and Shannon ended up in the same place (building something real, with a team that moved with them) but the paths looked nothing alike.
What I notice across both: the leaders who navigate an AI pivot well don't wait for permission or consensus before moving. They create clarity rather than asking for it. They treat team resistance as useful data, not an obstacle to route around. And they get in front of customers faster than feels necessary.
The trap most product leaders fall into is treating this as a resourcing or roadmap problem. Which features do we cut? Who do we hire? What does the 90-day plan look like? Those are real questions. But they're downstream of the harder one: how do you lead a team through a moment that nobody fully understands yet?
Lauren's answer was to reduce her own uncertainty by getting closer to the work. Shannon's answer was to own the decision clearly enough that her team could stop debating whether they were doing it and start debating how.
Both worked. And both required the same underlying thing: being willing to lead before you have all the answers.
If the pivot is coming from above and you need to bring your team through the shock, Lauren's path fits. Create clarity, get your hands dirty, get in front of customers.
If the conviction is yours and you're trying to get others to move with you, Shannon's path fits. Own the decision, use the resistance, find external expertise fast.
In both cases, the turning point wasn't a new framework or a better roadmap. It was the leaders deciding to stop managing the situation from a distance.
Lauren embedded in alpha customer calls. Shannon used her own product until she was annoyed enough to demand something better. Both got proximate to the problem.
There's a specific kind of clarity that only comes from direct contact: with the product, with customers, with the team's actual concerns. You can't synthesize your way to it. And in a fast-moving AI pivot, the leaders who get there first set the direction for everyone else.
An AI pivot isn't really an AI problem. It's a trust problem.
Your team trusts that the work they've done matters. Your customers trust that you're building for their real problems. And leadership trusts that you can move with speed and judgment at the same time.
The product leaders who navigate this well are the ones who hold all three simultaneously, not by having all the answers, but by staying close enough to the work to keep asking the right questions.
As Shannon put it: "There's still hesitation and there's still difficulty. But I would say in the product team, there's a lot more openness now. A lot more curiosity about what can be and what should be."
That curiosity is the thing you're building toward. Not just an AI feature. A team that's learned how to move through uncertainty together.
Want to hear the full conversation? Listen to Episode 5 of the Product Leaders Lab wherever you get your podcasts:
