.png)
A conversation with Chief Product Officer Ryan Bergstrom on multi-platform HCM, human-in-the-loop AI, and why expertise may become even more valuable in the age of agents.
One of the most interesting trends in HR technology right now is the rise of managed services. Every HCM vendor seems to be launching an outsourcing layer to sell over their technology stack.
Software companies are building managed services businesses. HR outsourcing firms are investing more heavily in software. Everyone is trying to figure out how AI fits into the equation.
During my conversation with Paychex Chief Product Officer Ryan Bergstrom, I found myself thinking about how differently Paychex enters that conversation.
This ‘new’ trend is something Paychex has been working on for years. The company has spent decades serving businesses across vastly different sizes and levels of complexity. It has built large service organizations around payroll, HR, retirement, and compliance. It has learned how to operate at scale in some of the highest-stakes workflows inside a business.
As we move into the AI era, this unique starting point creates a different set of opportunities for Paychex.
Rather than asking how to build services, Paychex is asking how AI can make existing services more valuable. Rather than debating whether software should assist users or take action on their behalf, Paychex is pursuing both simultaneously. And rather than forcing every customer into a single experience, the company is continuing to invest in a multi-platform strategy built for different stages of business growth.
Those themes surfaced repeatedly throughout my conversation with Ryan and offer an interesting look at how one of the industry's largest providers is thinking about the next chapter of HCM.
The HR technology market has increasingly split into two camps.
Some vendors believe the future belongs to a single platform that can serve every customer from startup to enterprise. Others believe different customer segments benefit from different experiences.
Paychex falls into the latter category.
Ryan pointed out that the company serves organizations ranging from a handful of employees to businesses with thousands of workers. While those organizations may need many of the same core capabilities, they often consume them very differently.
"It can come down to the complexity of their business, the stage that they happen to be in, what level of service and assistance they want from the buying cycle all the way to how they partner with us."
That reality creates a different product challenge than many software companies face.
The goal is to create experiences that fit different customer needs while still benefiting from the scale and resources of a shared infrastructure.
Ryan described the relationship between the platforms as "drafting off each other." While customers may experience different products, many of the underlying investments are shared. Tax engines, expertise, AI capabilities, and core infrastructure can benefit multiple platforms simultaneously, allowing innovations in one area to accelerate progress elsewhere.
One example is WISE, Paychex's new AI platform. Whether customers are using Paychex Flex or Paycor, they benefit from the same underlying intelligence and expertise powering the experience.
The result is a strategy that prioritizes customer fit without sacrificing the advantages of scale.
The conversation naturally shifted to WISE and how Paychex is thinking about AI more broadly.
Ryan described the platform through three stages: intelligence, assistance, and action.
The first layer focuses on awareness. The system understands the context of the business, identifies patterns, and surfaces recommendations based on what is happening inside the organization.
The next layer - assistance - helps users navigate workflows and complete tasks more efficiently.
Eventually, the platform moves into action.
"This is the point now of, you know, we reach the point where [you ask], ‘Does a person need to process payroll every single week, or can the system have payroll ready for them to simply review and approve the few changes that need to be made?"
What stood out to me was the intentional sequencing of the AI strategy.
Paychex serves hundreds of thousands of businesses. Some customers are eager to embrace automation. Others remain cautious about allowing AI to participate in critical workflows. Building AI at Paychex’ scale requires more than creating capabilities. You also need to help customers become comfortable with those capabilities over time.
Ryan described a future where users don't necessarily need to choose between doing work with AI or without AI. Instead, the platform continuously improves workflows in ways that feel natural and incremental.
"We're slowly bringing all of the changes into the system, all of the enhancements and improvements."
The emphasis is on helping work gradually become easier.
Of course, payroll and HR present a unique challenge.
Much of the excitement surrounding AI has been built around systems that generate content, summarize information, or answer questions. HR systems operate in a much higher stakes environment.
Payroll calculations, tax filings, employee records, benefits administration, and compliance obligations all carry real-world consequences.
"The impact of getting it wrong is real. We're impacting real people, real employees."
That reality creates an interesting tension.
Traditional HR software is largely deterministic. Users enter information, the system applies rules, and outcomes are predictable.
AI introduces probabilistic behavior. Systems identify patterns, make recommendations, and increasingly participate in decisions and workflows.
Ryan's answer was not to eliminate human oversight.
Instead, he described AI as another participant inside the system.
An agent can identify issues, recommend actions, complete work, and prepare outcomes. But it still operates within the permissions granted by the user it represents.
"If an agent is recognizing changes that need to happen or actions that need to happen, it's gonna come check in with its user before it goes and actually takes those actions."
Rather than imagining AI replacing administrators or managers, Ryan described a future where AI behaves more like an additional member of the team. It helps complete work, surfaces opportunities, and increases capacity, while humans remain responsible for review, approval, and accountability.
For categories as sensitive as payroll and HR, that balance may become one of the most important design decisions software companies make.
One of the most interesting parts of the conversation centered around managed services.
Over the last several years, many HCM vendors have expanded into HR outsourcing and developed advisory & compliance offerings.
The motivation is easy to understand: expand wallet share.
The challenge is that services businesses are difficult to build.
They require operational rigor, domain expertise, and the ability to consistently deliver outcomes in areas where mistakes can have serious consequences. Historically, many software vendors chose to let independent consultants and service providers deliver these services.
Paychex took a different path.
Long before managed services became a major strategic initiative across the broader HCM market, Paychex had already built significant businesses around payroll services, HR outsourcing, retirement administration, and compliance support.
What's particularly interesting is Ryan's belief that AI may actually strengthen the value of those services.
As automation handles more routine administrative work, customers and Paychex’ internal teams gain more time to focus on higher-level questions. Those questions often require expertise, context, and judgment rather than additional software functionality.
"We're actually finding if it brings the human more into the loop in a lot of these processes."
Ryan described a future where AI handles more of the repetitive work while allowing both customers and experts to spend more time discussing workforce strategy, compliance decisions, benefits planning, and organizational growth.
That perspective runs counter to many of the assumptions surrounding AI.
The common narrative is that automation reduces the need for human involvement. Ryan's view is that automation may actually increase demand for the highest-value forms of human expertise.
For a company that has spent decades building service organizations alongside software platforms, that could become a meaningful advantage.
Another theme that surfaced throughout the discussion was the changing relationship between software vendors and customers.
Historically, software pricing has largely revolved around access. Customers purchased licenses, subscriptions, or seats and then largely self-determined how effectively they used the platform.
Ryan believes AI creates opportunities for a different model.
"If AI is going to drive and actually be more of the digital workforce, we should be able to tell you as a customer very clearly, these are the things that it's doing for you. This is the value that you're capturing as a result of it."
As software becomes capable of completing work rather than simply facilitating it, customers increasingly evaluate platforms based on outcomes rather than functionality.
The industry is still figuring out what that evolution looks like commercially. Customers want predictable costs. Vendors want pricing models that reflect the value being created.
But the direction is increasingly shifting from buying software to buying completed work.
Toward the end of our discussion, Ryan reflected on what excites him most about the future.
Interestingly, his answer focused less on AI itself and more on the ability to solve customer problems faster and more effectively.
"We've spent a ton of time as we've come together refocusing on our customer in particular."
The discussion centered on helping businesses operate more effectively, helping users complete work more efficiently, and helping experts spend more time where their expertise matters most.
That mindset feels shaped by Paychex's position in the market.
A company that has spent decades running payroll, supporting HR teams, administering benefits, and advising small businesses naturally approaches AI differently than a company encountering those workflows for the first time.
The technology is changing quickly. The questions Paychex is trying to answer are surprisingly familiar: How do we help customers get better outcomes? How do we reduce complexity? How do we make expertise more accessible?
The tools may be new. The mission is not.
