CPO Rising Series: Paychex CPO on Reimagining The User Experience With AI

What happens when a 50-year-old software company tries to move at the speed of AI, and what does it really take to merge two very different product cultures after an acquisition? In this episode of the CPO Rising series hosted by Products That Count Resident CPO Renee Niemi, Paychex CPO Ryan Bergstrom will be speaking on reimagining the user experience with AI, how his team learned the hard way that accelerating only engineering just moves bottlenecks around, and why utilization is the only product metric that ultimately matters. He also gets into the HIPAM budget framework he uses to track and shift R&D spend, and why incumbents with deep customer data have a genuine moat in the AI era.

Join us for new conversations with leading product executives every week. Roll through the highlights of this week’s event below, then head on over to our Events page to see which product leaders will be joining us next week.

Show Notes:

1. Ryan started his career answering customer support calls — the “one 800 please help me” line — and credits that experience with building the customer empathy that shapes how he leads product today. Some of the best product managers he has worked with came from service roles, not traditional PM pipelines.
2. The Paycor-Paychex integration brought together two cultures that had each been genuinely successful — one more planful, one more iterative — neither of which was the right model for the combined organization going forward. The work was not picking a winner but finding what each side did best and building something new from both.
3. The most common form of resistance to AI adoption is fear about job security, and Ryan believes the most effective response is honesty and reframing. Paychex is investing in AI tools to invest in its people — to make them better, more capable, and more valuable — not to replace them.
4. Because nobody is truly an expert in how AI will change their role even 12 months from now, Ryan’s approach has been to encourage broad experimentation while providing training, guidelines, and clear expectations so people have enough structure to get started rather than feeling paralyzed.
5. Failure is not a word Ryan uses. What he focuses on instead is the time to learning — how quickly can you figure out that this path is not right, make a small adjustment, and get back on track. Reframing the concept from failure to fast learning changes how teams approach risk.
6. For AI-centered projects, Ryan’s teams demo every Friday. That weekly cadence is his favorite part of the week and it virtually eliminates drift because feedback comes before anyone can go far in the wrong direction.
7. Ryan uses a budget framework called HIPAM — Health, Innovation, Parity, and Modernization — to categorize every single thing his R&D organization is working on. Mapping spend across those four buckets makes it immediately visible whether you are under-investing in innovation or spending too much keeping the lights on.
8. The goal of the HIPAM framework is to drive progressively more spend into the innovation bucket each year. The way to fund that shift is not more budget — it is making the health and modernization work faster and cheaper, so the savings can be redirected.
9. Product is everyone’s lane in a software company. Ryan publishes the roadmap internally and presents it to all executive stakeholders every month — including what has changed and what they are explicitly not doing and why.
10. Explaining what you are not doing is just as important as explaining what you are. Every product organization has more requests than capacity, and the most effective way to hold that line is to connect every no to the company’s strategic pillars so it is a strategic answer, not an arbitrary one.
11. The biggest early mistake Paychex made in its AI rollout was going hardest on engineering tooling first. Engineers got faster, but product and design did not keep pace, so the bottleneck just moved upstream. The fix was not to go fix each bottleneck in sequence — it was to change the entire process.
12. Ryan calls his new approach “unpointed agile” — or jokingly “pointless agile.” The team has moved away from heavy estimation and backlog grooming toward a Kanban-style workflow where product, design, and engineering work hand in glove throughout the process rather than in sequential handoffs.
13. AI is not a bolt-on feature or a module within existing software — it is a full reimagining of the user experience. The starting point for any workflow is no longer necessarily the application itself, and users are no longer only humans. Agents interacting with other agents is now a real design consideration.
14. With a customer base as large as Paychex’s, the adopter curve for AI features is extremely wide. The people on the skeptical end of that curve will not give a second chance if the first experience does not deliver clear value, which means rollout pacing and quality control for AI features are higher-stakes decisions than for traditional features.
15. Utilization is the only product metric that ultimately matters at the team level. If people are not using what you built, you either built the wrong thing or you built the right thing badly. The job is not done when the feature ships — it is done when customers are meaningfully using it.
16. At Paycor, Ryan’s team doubled revenue per unit by rapidly expanding the product to help the company move upmarket. The growth was not about adding features to tick RFP boxes — it was about adding functionality that customers actually used, which in turn drove measurable business outcomes like the 11% better employee retention they saw in customers using talent management features.
17. The CEO of Paychex would say Ryan’s biggest contribution is driving a faster rhythm in the product organization and centering the company’s future around AI. Speed of delivery is the alpha effect — if product is not leading the pace, the business risks being disrupted not just by competitors but by new entrants who are not carrying legacy overhead.
18. Incumbents do have genuine moats in the AI era, and Paychex’s are the depth of customer data they hold and the trust they have already earned. AI-native startups cannot replicate years of proprietary data and customer relationships overnight, but incumbents have to actively use those advantages or they will erode.
19. Ryan uses AI to collapse the time between idea and alignment. He recently used Gemini to generate Amazon-style press releases and FAQ documents for new projects — a process that would have taken days — and the output sparked more focused conversations with his teams faster than traditional requirements documents would have.
20. The waterfall-to-agile transition took years. The agile-to-AI transition is happening faster, and product leaders who treat it as a tooling upgrade rather than a fundamental shift in how teams work will fall behind. The opportunity is not to make existing processes faster — it is to ask whether those processes still make sense at all.