I’m seeing more and more founders come back to run their pre-AI B2B companies.
Not because the board panicked over a bad quarter. Because it’s The Last Stand.
The pattern shows up at companies big enough to have a base that sustains them and not agentic enough to grow. In my own portfolio and in the public markets, that’s a lot of B2B companies at $50M+ ARR right now. Big enough to survive. Too slow to matter.
Not everyone needs one. If you sell consumption-priced infrastructure, AI arrived as volume on the meter. Twilio grew 22% last quarter, 17% organic, and raised its full-year growth guide by four points, all under a hired CEO. The Last Stand is for companies where AI showed up as a headwind instead, which mostly means seat-priced applications.
And at those companies, the founders are coming back because nobody else will do what it takes.
Screenshot
Daniel Dines Was Gone Four Months
The first one might have been Daniel Dines at UiPath.
Dines co-founded UiPath in 2005 and ran it as CEO or co-CEO through January 31, 2024, when he handed the job to Rob Enslin and moved to Chief Innovation Officer and Executive Chairman. Enslin was sole CEO starting February 1.
On May 29, 2024, UiPath announced Enslin was resigning effective June 1 and Dines was being re-appointed CEO the same day.
Four months.
The quarter they reported that day tells you why. ARR was $1.508 billion, up 21%. Revenue was $335 million, up 16%. And they cut guidance for both the quarter ahead and the full year. The stock, which had touched nearly $27 in February, dropped about 35% to under $12.
AI hit RPA as fast as it hit any category in B2B. Dines was back before the second full quarter of the transition closed.
Aneel Bhusri Came Back to a Company Guiding to 11% Growth
Aneel Bhusri did the same thing at Workday on February 9, 2026.
Carl Eschenbach joined as co-CEO in December 2022, became sole CEO in February 2024, and was out two years later. Workday confirmed to TechCrunch that Bhusri’s return is permanent, not a placeholder during a search.
The framing was AI, in Bhusri’s own words: AI is a bigger transformation than SaaS, and it will define the next generation of market leaders.
The numbers behind that sentence: Workday’s stock was down about 15% in 2026 before August and sat more than 40% below its 2024 peak. Q2 FY27, reported August 27, put subscription revenue at $2.471 billion, up 13.9%. Management guided fiscal 2028 subscription growth to roughly 11%.
That’s the shape of the problem. A $10 billion revenue business with 97% gross retention, growing at 13% and decelerating. There’s nothing wrong with it. It just isn’t going anywhere.
Intercom Founder Eoghan McCabe Returned as CEO, Renamed a 15-Year-Old Company After Its Agent
Eoghan McCabe stepped out of the Intercom CEO seat in 2020 and came back at the end of 2022, a few weeks before ChatGPT launched. His own account of the pivot: in need of a reboot, they jumped on weeks-old LLMs and built the customer agent category.
Then in May 2026 he did the thing a hired CEO almost never gets to do. He renamed the company Fin, after the agent. Fifteen years of brand, the chat bubble in the corner of half the internet, gone in favor of the new product.
Weeks later, on June 15, Salesforce agreed to buy Fin for about $3.6 billion, with roughly 30,000 customers and $400M+ in recurring revenue.
It’s the best outcome of the three, and the multiple is the part most people skip. $3.6 billion on $400M+ of recurring revenue is under 9x. Sierra and Decagon raise at AI multiples. The company that actually executed the rebuild got a B2B multiple. Coming back and pulling it off got McCabe a real exit, not a re-rating.
Dines Rebooted UiPath. Four Straight Profitable Quarters, and Still 12% Growth
UiPath reported Q2 FY27 on September 3, 2026. Dines has been back 27 months. The results:
- ARR $1.938 billion, up 12%. It was up 21% the quarter he returned.
- Revenue $410 million, up 13%, or 16% normalized for an $8 million FX headwind.
- Net new ARR $37 million in the quarter.
- Net revenue retention 109%. It was 110% at the start of FY26.
- Non-GAAP operating margin 22%, up from 17%.
- Fourth consecutive quarter of GAAP operating profit, $32 million against a $20 million loss a year earlier.
- Stock comp $45 million, down 42% year over year, now 11% of revenue.
- Customers at $1M+ ARR up 21% to 387, while smaller customers drove most of the attrition.
Dines’ own line on the call: they spent the past two years transforming the platform and strengthening execution.
He rebooted the company. A business that was cutting guidance and losing money in mid-2024 is now GAAP profitable four quarters running, expanding margin, holding 97% gross retention, and growing its $1M+ customer count 21%. It is still growing, in double digits, in a category AI hit head-on. There was no version of that outcome without him in the chair.
What it has not produced yet is reacceleration. The growth rate is lower than the day he walked back in. That’s the honest read across all three of these: the founder coming back is what stops the decline and rebuilds the foundation. Restarting the growth is a separate job, and it takes longer than anyone wants.
Eschenbach Delivered Exactly What He Was Hired For
Non-founder CEOs, in my count, 9.5 times out of 10 can’t pull a pre-AI B2B company out of a slow death spiral. That’s my read from the boards I sit on, not a dataset. But the mechanism is easy to see, and Workday’s own press release states it out loud.
The company credited Eschenbach with global growth, an expanded industry focus and strengthened operational discipline. He did all of that. He was hired to do all of that. He delivered it.
None of it was the rebuild.
A hired CEO arrives with a plan, a comp package tied to that plan, and a board that approved it. The AI rebuild breaks every line of that plan. It means cannibalizing your own pricing model before a competitor does. Eating margin for six to eight quarters. Killing products that still bill. Telling public investors the number goes down before it goes up.
A professional CEO who does that can get fired for it, and doesn’t get thanked for it either way. A founder who does it is doing what he did in year one.
I watched the mild version of this myself. Adobe killed EchoSign’s free edition within weeks of my leaving. Nobody was being malicious. The next person just cares less about the thing you bled for.
Bhusri and Duffield Control 68% of the Vote
The Last Stand requires a founder who can actually come back. That’s a governance question, and most companies answered it years ago without realizing it.
At Workday, Bhusri and Duffield control 68% of the vote. Bhusri never left the building; he was executive chair. Dines was Chief Innovation Officer and Executive Chairman when he came back, which is why UiPath’s transition took four months and not a nine-month search.
McCabe was the hard case. He was fully out for two years, and it took board members asking him to come back.
If your founder is off the cap table, off the board, and out of the company, you don’t have this option. You have a search committee.
Silver Lake Had Been in Talks for Months Before Anyone Heard
If nobody makes the Last Stand, somebody else makes it for you.
On August 13, 2026, Reuters reported Silver Lake had been in talks for months about taking Workday private. The stock rose about 18%, its best day since 2016, lifting the market cap from roughly $43 billion to nearly $51 billion. No price, no terms, no assurance of a deal.
Workday added $8 billion of value on the possibility of leaving the public markets. Investors were saying the rebuild is more likely to happen without quarterly guidance attached to it.
PE is the other answer to the same question. It’s the one that arrives when the founder isn’t there or won’t do it. Bhusri, seven months back in the seat, is now running an AI rebuild while a buyer circles.
The rebuild is showing up in his numbers: AI drove more than $100 million of new ACV in Q2, over 25% of the total; AI SKU ARR is around $600 million, up more than 200%; 5,500+ customers are using at least one agent; and a consumption-based credit model is coming. That’s the most tangible progress of the three companies here. It is also seven months of work against a decade of seat-based revenue, and the growth guide still says 11%.
Dines Had to Go Back at 21% Growth
That’s the whole point. UiPath was growing 21% and generating cash and had a founder sitting right there as Executive Chairman with a hand-picked CEO running it, and it still took four months for him to conclude he had to take the job back himself.
If your pre-AI company is grinding toward 10% growth and the founder is watching from the board seat, the window to come back is not the day the number goes negative. Dines came back at 21%.
What the return buys is real: the decline stops, the company gets rebooted, growth holds, and you keep playing. Reacceleration is the part none of them have delivered yet.
It’s not too late. But it’s getting too late. And you really, really need your founders for the Last Stand. No one else is going to do what it takes
