Workhelix Leverages Years of Research to Guide Enterprises in AI Application

AI has the potential to revolutionize how we work, but figuring out how to actually get value from AI isn't as simple as just applying any AI tool to any task. For big companies, it can be tough to sort out which AI applications will really boost their business and which ones are just hype. That's where Workhelix comes in—they're trying to solve this puzzle.

Workhelix is a tech startup that teams up with enterprises to help them better understand and keep tabs on their AI automation. They dive deep into a company's employee roles, breaking them down into specific tasks, and then they rate each task on how well it might work with AI. This approach helps companies create a roadmap for where and how to use AI, and it lets them keep an eye on whether the AI they've implemented is actually doing its job.

James Milin, the co-founder and CEO of Workhelix, told TechCrunch that a lot of companies mess up AI adoption because they try to apply it to entire divisions, which is way too broad to see real benefits.

"That's not a systematic, rigorous way to adopt generative AI and is part of the reason people are often so let down," Milin said. "But if you take a closer look at all the jobs in an organization, break them down into specific tasks, and then score each task on its potential to be enhanced by generative AI, you can develop a really solid, quantitative approach to adopting it."

Workhelix's method of breaking down roles into tasks is built on years of research into how technology affects productivity, led by Erik Brynjolfsson, who's the director of Stanford's Digital Economy Lab and one of Workhelix's co-founders.

"In a lot of our work, there are tasks where machines don't help much," Brynjolfsson said. "Humans need to be involved. But then there are other tasks where machines are super helpful. And pretty much every project we look at has a mix of both."

Brynjolfsson has been diving into this divide between technology and productivity for over a decade. Before Workhelix, he was sharing his research and methods through papers and talks in boardrooms, but he realized that adding a software component could help reach more companies.

Brynjolfsson, who's also the co-chairman of Workhelix, teamed up with Andrew McAfee, the co-director of MIT's initiative on the digital economy and one of Brynjolfsson's co-authors; Daniel Rock, a Wharton professor; and Milin to launch Workhelix in 2022.

The company rolled out its product in April 2024 and has seen a strong demand from enterprise customers like Accenture, Wayfair, and Coursera, among others. Workhelix's first dozen enterprise clients came on board without any paid advertising, Milin said.

"This is something companies are really eager for," Brynjolfsson said. "They haven't seen anything like it before. Sure, there are consultants out there, but they don't have tools like ours. We're filling a huge gap in the market."

Workhelix recently raised $15 million in a Series A round led by AIX Ventures, with participation from Andrew Ng's AI Fund, Accenture Ventures, and Bloomberg Beta, among other VCs. They also got funding from angel investors like LinkedIn co-founder Reid Hoffman, OpenAI co-founder Mira Murati, and Jeff Dean, the chief scientist at Google DeepMind and Google Research, among others.

Shaun Johnson, a founding partner at AIX Ventures, told TechCrunch that he was introduced to the company through Brynjolfsson's work at Stanford; one of AIX Ventures' investing partners, Christopher Manning, is the director of Stanford's artificial intelligence laboratory. Johnson said he got the pain point Workhelix was addressing right away.

"Erik, Andy, and Daniel have amazing access to the Fortune 500 C-suite and to customers," Johnson said. "It's a perfect match between founders and market, and their approach is a perfect match between founders and product. That's what made us want to dive in."

Workhelix plans to use the new funds to expand the number of tasks and KPIs their software tracks. They'll also keep building up the internal tools for the data scientists who directly help enterprise customers alongside Workhelix's product.

In today's market, where everyone's obsessed with speed and automation, it's interesting that Workhelix's business model isn't just about software—it also includes a human touch. The company sticks by this approach, even though it makes scaling up a bit trickier. That's because, as Milin said, Workhelix wouldn't be as effective if it were just another software platform.

"I think there's a trillion-dollar opportunity here to create value," Brynjolfsson said. "We're not going to capture all, or even most of it, but we want to unlock it. As James mentioned earlier, this is the biggest technological revolution that's ever happened, and very few people are thinking about unlocking the business side of it."