What happens when you empower 1,800 employees to shape your AI strategy?
CEOs are going into 2027 nervous, with revenue confidence at a five-year low (PwC). AI was supposed to help, and for most companies it hasn’t yet. A Google Cloud study released in late September found more than 8 in 10 executives see value in AI, but only about a quarter are getting a real return, and PwC found the top 20% of large companies capture 74% of AI’s financial gains (IT-Online).
Much of that gap traces back to how AI gets rolled out. A central team picks the tools and use cases, then pushes them onto employees who had no say. One trait Google tied to the ROI leaders was decision-making authority, and in most top-down rollouts, the people doing the actual work have none.
Klaviyo flipped the model. Leadership set the direction, then the 2,400-person marketing automation company asked every employee to find an AI use case worth building in their own job, build it, and demo it. Between the CEO’s kickoff memo in August 2025 and demos in December, 1,800 employees, about three quarters of the company, had something working.
In the quarters that followed, revenue per employee climbed 25 to 28% with headcount flat. Here’s how they pulled it off, and how a bottom-up program can sit alongside the top-down AI work most organizations already have underway.
It All Started with a CEO Memo
In August 2025, Klaviyo’s CEO told the whole company that AI was here and they could either jump in or fall behind. What followed was a roughly five-month push with three major milestones:
- Explore, Play, and Learn – Klaviyo ran an Enablement Month with demos from OpenAI, Google, Atlassian, Zoom, and local tech leaders. Everyone got a learning stipend to spend on whatever AI tools they wanted to try, work-related or not. The goal here was to bring people up to speed on AI to help them better understand what’s possible to ultimately connect it back to their role.
- Develop a Use Case and Build an MVP – Every employee had to pick one AI goal that would create value in their own role and spend three months building it. HR didn’t assign anything. Employees chose their own goals, and managers checked that each one was worth three months of effort before work started.
- Present a Demo – By December, every employee was expected to demo what they built. The challenge also counted in year-end reviews, which made it part of the job.
Blake Schuller, then Klaviyo’s AI Transformation Lead, ran the effort. Her logic was simple. Employees know their own work better than any central team does. “You know your role best,” she said.
The Impact of the Bottom-Up AI Initiative
The ideas and use cases spanned a wide breadth of categories. Some people built automations that save them hours every week. A recent college grad in finance taught herself HTML and CSS to build a site that pulls live stock data. Plenty of others tried something, hit a wall, and switched approaches. Klaviyo treated that as part of the process.
By the end of 2025, 1,800 of roughly 2,400 employees had demoed a working AI use case. That’s 75% of the company in one quarter, driven largely by pushing decisions down to the people closest to the work and giving them room to explore and create business value in a supportive environment.
Although we don’t have specifics on the overall ROI from this initiative, what the company has reported in its earnings since then gives us directional evidence. Leadership has tied these gains to Klaviyo’s internal “AI-First” push. Here are a few of the stats reported in May 2026, just 9 months after the program kicked off:
- Revenue Per Employee: Annualized revenue per employee topped $600,000, up more than 25% year over year.
- Code Deployment Speed: Engineers were shipping code at nearly double the rate of a year earlier.
- Operating Margin: Non-GAAP operating margin hit a record 16%.
6 Best Practices & Key Takeaways
There are six key takeaways that we see time and time again, and they apply whether your program is about growth, cost savings, or continuous improvement. None of these are rocket science, but navigating the change management to make them happen can be challenging, especially at this scale:
- Leadership set the expectation and tied it to reviews – The CEO memo explained why AI mattered before anyone was asked to build anything. Then Klaviyo made the challenge part of year-end performance reviews. That turned it from a nice-to-have into part of the job, and it told employees leadership was serious.
- People got room to explore, fail, and learn – Before the real ask, Klaviyo ran a month of demos from companies like OpenAI and Google and gave every employee a stipend to try any AI tool they wanted. Some use cases flopped later, and people who hit a wall and switched approaches weren’t penalized. Reward only the winners and people quietly stop trying.
- A time-boxed challenge, on top of the always-on stream – Open suggestion boxes tend to collect a trickle of ideas and then go quiet. Klaviyo gave everyone three months and a December demo date, which created urgency and a shared finish line. If you’re setting one up, our guide on how to create effective challenge statements covers how to frame it, and here’s more on running innovation challenges.
- Employees owned their ideas – Nobody was handed a use case. Each person picked a problem in their own job, the one they knew better than anyone else. Ownership is why 1,800 people built something real instead of checking a box. This is internal crowdsourcing in practice.
- Managers coached, and no central team became the bottleneck – This is the big one. Klaviyo pushed decisions down to managers, who confirmed each idea was worth the effort and guided it along. A central AI team can only evaluate so many ideas, so impact stalls at that team’s capacity. The research points the same way: Gallup found manager support is the strongest predictor of whether employees use AI, and Google’s ROI leaders stood out on decision-making authority. Giving autonomy to the people closest to the work is how you scale past a handful of pilots to 1,800 working use cases.
- A system to capture and track impact – Klaviyo added an AI Impact tag to performance reviews, ranging from no impact to changing how the business operates, and it now showcases employee wins in company meetings. Tracking 1,800 ideas in email threads or spreadsheets falls apart fast. You need one place where every idea is visible, reviewed, and measured. Our 2026 guide to capturing employee ideas walks through setting that up.
Klaviyo gave employees a clear challenge, real support, and ownership of their ideas and the team delivered at scale. The same approach can be applied beyond AI to any innovation or continuous improvement program you have live today. We hope this helped provide some starting inspiration for you and your team!
Ready to Unlock the Innovation Potential of Your Workforce?
Ideawake is a shockingly simple idea management platform. Our customers double the ROI they get from employee ideas and spend half the time managing them. Teams use it to run time-boxed challenges, route ideas to managers for review, track impact, and share wins from one place. Our benefits and testimonials page has customers explaining it in their own words.
Book a demo with Ideawake to see how a company-wide idea challenge would run at your organization. If you’re not ready for a demo, start with our free AI Idea Campaign Playbook to plan your own company-wide AI challenge.
Sources
- How Klaviyo helped 1,800 employees find a great AI use case, Section, Feb. 12, 2026
- PwC’s global chairman on the 29th CEO Survey, Fortune, Jan. 19, 2026
- Global Employee Engagement Continues Decline, Gallup, April 7, 2026
- Execs see AI value, but ROI is still elusive, IT-Online, Sept. 28, 2026
- Klaviyo Q4 Earnings Call Highlights, MarketBeat via Yahoo Finance, Feb. 11, 2026
- Klaviyo Q1 2026 Results, Form 8-K, SEC, May 5, 2026
- Klaviyo (KVYO) Q2 2026 Earnings Call Transcript, The Motley Fool, Aug. 12, 2026