How to Build an AI-First Culture with 1,300 People. Real Story.
The Bold Experiment That Changed Everything
What if your company of over a thousand employees and hundreds of company clients stopped chasing quarterly goals for an entire month and instead invested in making every single employee an AI builder? That's exactly what AppsFlyer did, and the results reveal a fundamentally different approach to AI transformation that most companies are missing.
In a recent conversation on the Deconstructor of Fun podcast, Barak Witkowsky, Chief Product Officer of AppsFlyer, shared how the 1,300-person company took an unprecedented step: they paused regular business objectives and put everyone through a four-week AI builder course. Not as something extra on top of their work, but as the work itself.
"We're pretty much all in on AI," Barak explained. "What we try to achieve is that each of our 1,300 people will know AI better than management so we can leverage the wisdom of the crowd."
The Problem with Traditional AI Adoption
Most companies approach AI transformation with a top-down mandate. Leadership sets KPIs, defines processes, instructs their teams to "hire less and employ more AI” and expects teams to figure it out. The problem? This creates what I call the "AI hesitancy gap."
Talk to executives at major companies, and they're bullish on AI. They see the potential, read the headlines, and push their teams to "do more with AI." But talk to directors, managers, and individual contributors, the people who actually execute, and you'll find something very different: uncertainty, uneasiness, and often a quiet dread that they're about to become obsolete.
"Everyone has some imposter syndrome, thinking all others are working so great with AI and not me," Barak noted. This creates a paradox where everyone is simultaneously pushing for AI adoption while feeling inadequate about their own progress.
AppsFlyer's Dual Approach: Top-Down Meets Bottom-Up
What makes AppsFlyer's transformation remarkable is how they combined strategic direction with grassroots innovation. Yes, they set KPIs and defined processes that must become AI-driven (they call these "ATOMs"). But the secret sauce was democratizing AI knowledge across the entire organization.
The company ran hackathons, recognized champions, and most importantly, completely freed people's time to innovate with AI. As Barak put it: "My number one source of knowledge for AI is coming internally from our 1,300 people."
Think about that. The CPO of a major tech company doesn't rely primarily on external consultants, conferences, or vendor pitches. His best insights come from his own team, who today tend to know more about AI tools than developers at other companies.
This investment came directly from CEO Oren Kaniel two years ago.
"When he first talked about it, I felt like he's almost too much all in," Barak admitted. "But now I don’t meet anyone who doesn't think anymore that AI is not gonna shape everything."
From Attribution Platform to Modern Marketing Cloud
AppsFlyer's AI transformation wasn't just internal. The company simultaneously evolved its product from an attribution platform into what they call a "modern marketing cloud," launching a suite of AI agents designed to automate marketing workflows.
Three market forces drove this evolution:
CFO scrutiny: Marketing teams face intense pressure to prove business impact, especially in today's economic climate
Omnichannel complexity: The average American family now has 17 connected devices, forcing marketers to break down organizational silos
AI disruption: Marketers expect their tools to help them navigate the AI era, not just measure campaigns
In their Fall release, AppsFlyer introduced AI agents that function as new team members. Think of them as creative opportunity spotters, audience builders, and strategic analysts who work alongside human marketers.
But here's the critical insight: AppsFlyer is not just building AI products. They're helping their customers become AI-first companies.
Building Customer Confidence
One of the most fascinating parts of our conversation was about confidence. Specifically, how do you get marketers to trust AI-driven insights when they're deploying tens of millions of dollars in ad spend?
Barak's answer was refreshingly honest: "It's hard for us to measure if the confidence of the customer is growing." Instead of chasing vanity metrics, they focus on behavior: Are customers making bolder, faster decisions? Are they willing to shift $20 million based on measurement data? Are they cutting campaigns after two days when the numbers don't work?
"If they're doing that all over the world, across different verticals, this is honestly my best compass to know that they can trust and have that confidence," Barak explained.
This extends to their AI agents as well. AppsFlyer is working hand-in-hand with major brands to build trust in autonomous marketing, step by step, much like you would with new team members.
What This Means for Performance Marketers
"We have 100% confidence that we're moving towards an era of autonomous marketing," Barak said. So what happens to performance marketers when performance marketing becomes automated?
Rather than dystopian job loss, Barak sees an optimistic future:
"Every marketer is a boss of agents, which are bosses of other agents. Dream with it. Build whatever you want with it. Take marketing to the extreme."
The key insight: as AI makes building products easier, marketing becomes more crucial, not less. When anyone can build an app, the bottleneck becomes user attention, which is marketing's domain.
Marketers who learn to leverage AI as an extension of their capabilities, bringing creativity, strategy, and business context while letting agents handle execution, will be more dominant than ever. Those who don't? Their future is less certain.
Philosophical Musing: AI Done Right
AppsFlyer's transformation offers a blueprint that flies in the face of conventional wisdom. Instead of asking employees to learn AI "on top of" their regular work, they made learning AI the work. Instead of centralizing AI expertise, they democratized it. Instead of just using AI internally, they're helping their entire ecosystem adopt it.
The investment was significant. The ROI is still being measured. But the cultural shift is undeniable. When you invest in making every single person in your organization an AI builder, you don't just get productivity gains; you get 1,300 people innovating on your behalf.
As Barak put it:
"A company that is not making all their people excited about AI is probably not going to be here five to seven years from now. It's like in the year 2000, saying that you do not want to connect your company to the internet."
For companies still wondering how to approach AI transformation, AppsFlyer's example is clear: press pause on business as usual, invest deeply in your people, and build from the ground up. The quarterly goals will still be there. But your company, and your people, will be fundamentally different.
This post is based on a conversation with Barak Witkowsky, Chief Product Officer at AppsFlyer, discussing their company-wide AI transformation and the evolution of their product into a modern marketing cloud.
