Why isolved’s AI Opportunity Brought Michael Haske Back to HCM

Michael Haske had spent nearly 30 years in human capital management when he finally decided to leave the industry.

After helping build and scale several major HCM businesses, he moved into pure-play AI and assumed the HCM chapter of his career was finished.

“I finally got out, I'm not going back,” he recalled telling recruiters.

Then one of them called with an opportunity they knew he was going to resist.

It was isolved.

“I was like, damn it, I'm gonna have to like seriously consider coming back to HCM.”

What changed his mind was less about returning to HR technology and more about what he saw when he looked at isolved through an AI lens.

The AI opportunity was the reason to come back

Haske spent his time away from HCM at Krista AI, working with companies on how to deploy AI internally and build it into their products.

That experience changed how he thought about the technology.

“[AI] is a full-time job and a domain of expertise.”

His takeaway was that application software companies should be cautious about trying to build their own foundational AI capabilities when frontier model providers are investing at a scale few companies can match.

When Haske evaluated HCM companies, that created an interesting divide.

Smaller vendors could struggle to keep pace with the required investment. Larger vendors had resources, but many had already committed to AI strategies that Haske would have approached differently.

isolved sat somewhere in between.

It was already a $600 million-plus business with 200,000 customers and 9 million employees on record. It had the resources to invest, but its AI strategy was still early enough for Haske to apply what he had learned.

That combination pulled him back into the industry.

From helping people work to owning outcomes

Much of the first wave of AI in HCM has focused on making existing work easier.

Generate a job description. Summarize a document. Answer an employee question. Help an administrator find information.

Haske sees a much larger opportunity ahead.

“The real unlock comes when you can get AI to own outcomes.”

That means agents moving beyond assistance and becoming responsible for completing work across systems.

For that to happen, agents need access to the same tools and data that employees use today. An HR workflow rarely begins and ends inside the HRIS. It may touch the CRM, finance platform, email, or another part of the company’s technology stack.

That is part of the logic behind isolved’s partnership with Anthropic and its work around connectors. Haske also said the architecture is being designed so isolved can work with other models over time.

The goal is to give agents enough context and access to complete work without locking the platform into a single AI provider.

Legacy HCM can still compete in an AI-first world

AI-native startups have one obvious advantage. They get to build without decades of technical baggage.

But Haske does not think that makes established HCM platforms obsolete.

“You absolutely can do it with legacy solutions.”

The requirement is getting the underlying data into a structure AI can actually use.

“Garbage in is still garbage out.”

Established platforms may need to open up APIs, connect siloed data, and modernize parts of their architecture. In return, they have years of payroll, HR, workforce, and compliance data that a new entrant does not.

There is another constraint inside HCM: some workflows should remain deterministic.

“You can't replace the truly deterministic stuff, especially in the payroll HCM world, with the generative AI tool.”

Payroll calculations and compliance rules need predictable outputs. Generative AI can operate around those engines, helping users interpret information or take action while deterministic software continues handling areas where accuracy cannot be probabilistic.

Haske expects the interface to follow a similar hybrid model.

Some work will move naturally into conversations. Other tasks will remain faster through traditional screens and controls. isolved’s upcoming UI refresh is being designed with both experiences in mind.

AI could also strengthen isolved’s service model

The other surprise Haske found after joining isolved was the scale of the business itself.

“I've come to say it's the best kept secret in HCM.”

Beyond the customer base and product footprint, he pointed to isolved’s service organization and its network of more than 250 payroll partners as advantages he had underestimated from the outside.

That service model also shapes how he wants the company to use AI internally.

Instead of viewing customer support and implementation primarily as automation targets, Haske sees an opportunity to give employees better tools and shorten the long learning curve that comes with working in payroll and HCM.

A newer service employee could operate with knowledge that historically took years to accumulate. Implementation teams could get to answers faster. Support employees could spend less time searching for information.

It is a different AI equation than simply asking how much headcount software can replace.

Haske pointed to companies already seeing productivity gains from AI and said many are expanding rather than shrinking their teams.

“They are unlocking potential.”

That may ultimately be the larger bet behind isolved’s AI strategy.

The company already has the payroll infrastructure, customer data, service operation, and distribution footprint. Haske’s job is to layer a new generation of intelligence on top without losing the pieces that made the business work in the first place.

After trying to leave HCM behind, that was enough to bring him back.