Eightfold’s 10-Year CEO Reset: Why Ashutosh Garg Handed the Keys to Joe Korngiebel
Founders love saying they’ll hire a better CEO. Ashutosh Garg just did it after 10 years at Eightfold—and the timing tells you exactly where AI companies are heading.
Founders love telling investors they’ll eventually hire a better CEO. Most of them are lying to themselves.
Ashutosh Garg has done the harder thing: after a decade building Eightfold AI, he has handed the chief executive job to Joe Korngiebel and moved into the chairman role. It is not a sentimental handover. It is a blunt admission that the next phase of an AI company is not won by having the best original idea. It is won by turning that idea into software big companies will trust, buy, renew and build their operations around.
That is a very different job.
The handover is the story—not the press release
Eightfold announced on October 13, 2026 that Korngiebel is its new CEO, replacing co-founder Garg, who led the business for 10 years and remains on the board as chairman.
Korngiebel is not a celebrity founder brought in to spray buzzwords around. He is an enterprise-software operator. He spent 14 years at Workday, including a stint as chief technology officer from 2016 to 2020, and most recently served as executive vice president and chief strategy, product and technology officer at Dayforce.
That résumé matters because Eightfold’s challenge is no longer proving that AI can make hiring, internal mobility and workforce planning smarter. Plenty of buyers already accept that premise. The challenge is making the technology useful enough—and safe enough—to sit inside the machinery of a major employer.
That means integrations. Security reviews. Procurement committees. Change management. Product reliability. Sales cycles that can last longer than a founder’s patience. And, critically, getting a chief people officer, a chief information officer and a finance team to agree that the thing is worth paying for.
None of that is glamorous. It is where fortunes are made anyway.
Eightfold says new customers for its agentic products have quadrupled this year. Take that as a company claim, not independent proof of a gold rush. But it does explain the boardroom logic. When demand shifts from software that recommends things to software that can take actions, the leadership standard rises sharply. You are no longer selling a clever dashboard. You are asking customers to allow software into decisions about their people.
If that goes wrong, nobody cares how beautiful your demo was.
Ashutosh Garg is not leaving the field
The lazy read is that Garg has stepped aside because the founder chapter is over. That misses the point.
He remains chairman, and his influence is still obvious in the broader strategy around AI and work. Just before the CEO transition, Fortune reported on TwinCafe, a platform Garg launched that allows job candidates to create a digital twin from material such as a résumé, LinkedIn profile and work samples. The candidate can share the twin with a recruiter before a human interview takes place.
The pitch is straightforward: stop wasting the first 20 minutes of an interview reciting a résumé. Let the machine answer basic questions about work history, skills and career gaps. Save the live conversation for judgment, curiosity, communication and whether the person can actually think.
In theory, that is sensible. Most first-round interviews are not interviews. They are administrative theatre with better lighting.
Garg’s bigger idea is even more aggressive: a recruiter’s agent could eventually interview a candidate’s agent before either human gets involved. That could let companies assess a much wider pool of applicants than the handful a recruiter has time to screen manually.
But this is precisely why Eightfold needs a CEO with deep enterprise scar tissue.
A digital twin can make a candidate more legible. It can also make them more polished, more optimised and potentially less real. The moment candidates know their AI twin is being assessed by an employer’s AI agent, the arms race starts. Résumé keyword stuffing was amateur hour. Now people will tune their digital proxies to sound competent, credible and perfectly aligned with every role.
That is not a reason to reject the technology. It is a reason to build verification, guardrails and human accountability before the software becomes the gatekeeper.
AI hiring has a trust problem, not a capability problem
Every founder in AI is currently saying a version of the same thing: “Our product gives people time back.” Fair enough. A decent tool should.
But in hiring, time saved is not the only scorecard. The company also has to answer some awkward questions.
Can a candidate correct false assumptions made by the system? Can an employer explain why a person was rejected? Can the business verify that an AI-generated candidate profile reflects the individual rather than a professional ghostwriter with a good prompt? Who owns the data? How long is it stored? And when the system makes a stupid call—as every system eventually will—who has the authority to override it?
These are operator questions, not technology questions.
That is why Korngiebel’s appointment is more interesting than the usual CEO musical chairs. Eightfold is effectively putting an enterprise-product executive in charge at the moment its market is moving from intelligence to action.
There is a massive difference between software that says, “Here are 25 potentially suitable people,” and software that helps decide who gets an interview, who gets promoted, who is reskilled or who is overlooked. The latter affects livelihoods. It also creates a much higher bar for accuracy, transparency and customer confidence.
The companies that win this market will not necessarily have the flashiest model. They will be the ones whose customers can explain what the product does, where it stops, and who remains accountable.
Boring? Maybe. Valuable? Bloody oath.
The overlooked angle: this is founder maturity, not founder failure
There is a bizarre myth in startup land that the best founder must remain CEO forever. It is nonsense.
The skills required to start a company are not identical to the skills required to scale one. Early on, you need conviction before evidence, speed before process and the ability to persuade people to follow you into a fog bank. Later, you need systems, operating rhythm, executive depth and enough discipline to say no to the 19 ideas that would distract the business.
A founder who cannot recognise that transition is often the biggest risk on the cap table.
Garg’s move says something worth noticing: he is choosing to keep strategic influence while putting day-to-day execution in the hands of a person built for enterprise scale. That arrangement can fail, of course. Founder-chairman relationships go pear-shaped when nobody is clear about who owns the final call.
The incoming CEO cannot be a caretaker. The chairman cannot be a shadow CEO who drops into product reviews, reverses decisions and confuses the team about who is actually in charge.
For this to work, Korngiebel needs full operating authority. Garg needs a defined chairman role: strategy, customer relationships, recruiting top talent, capital and long-term product vision. If the division is fuzzy, staff will play politics and decisions will slow down.
That is the trap. Good governance is not about adding more smart people to a meeting. It is about making sure everyone knows who gets to decide.
What this means for you
If you are a founder, do a brutally honest audit this week. Ask yourself: am I still the best person for the company’s next two years, or am I simply the person who started it?
Do not ask your mates. Ask your best executives, your hardest-nosed board member and your most important customers. Separately. Then listen for the repeated answer.
If you are hiring a senior executive, stop hiring for biography alone. Hire for the bottleneck in front of you. Need product-market fit? Hire a builder. Need repeatable revenue? Hire someone who has carried a quota and lived through enterprise procurement. Need global scale? Hire an operator who has run complex systems when they break at 2 a.m.
And if you are introducing AI into hiring, promotions or performance management, use this rule: automate the admin, not the accountability.
Let AI summarise. Let it organise. Let it surface candidates and spot patterns humans might miss. But keep a named human responsible for any decision that materially affects a person’s job, pay or future.
That is not anti-AI. It is pro-business.
The companies that treat AI as a way to dodge judgment will create faster, more expensive mistakes. The companies that use it to make good people sharper will build an edge that lasts.
Eightfold’s CEO handover is a reminder that scaling AI is not mainly a model problem. It is a leadership problem. And the founders who understand that early will keep more of what they built.