How AI Startups Are Redefining Company Culture to Retain Top Talent

In the AI era, speed isn’t just a competitive advantage: we believe it’s a survival requirement. But traveling at speed creates many challenges when building a sustainable company culture that attracts and retains high-performing people. The AI startups pulling ahead are cultivating culture from day one, knowing it can be the difference between scaling and failing.

As the operating team for EQT’s early-stage investing arms – Ventures and Growth – we’re close observers of the AI startup ecosystem. In our last article, we looked at how many leading AI startups are rethinking hiring – abandoning traditional talent profiles in favor of founder energy, infinite skills and conviction.

Getting the right people through the door is only half the battle, however. The other half is creating a culture – with processes and systems – to ensure this new talent thrives. While there’s no one playbook for startup culture, we’ve been exploring how some firms bring structure to their teams of often highly self-directed individuals.

“Exceptional people are really a pain in the ass to manage once you have them on board,” says Olivia Elf, chief of staff at Sana Labs, an EQT-backed AI startup focused on knowledge work. “They expect and demand a lot from you as a company and hold you accountable all the time to help them grow and to challenge them.”

Handling promotions

Sana, which earned a reputation for hiring exceptionally talented individuals and for its culture “rituals,” helps talent grow by promoting the vast majority of its managers from within. Leadership at Sana comes from earning trust within the organization, not from traditional markers like tenure or previous experience. Employees earn “sensei” status if their peers start seeking them out for their knowledge or ask to be managed by them.

“We didn’t write this playbook and then hope that the culture would follow, we did it the other way around,” Elf says. “We picked up on the things that we saw were already lived in the organization, and then we articulated them.”

Meanwhile, AI voice startup ElevenLabs recently scrapped titles entirely in favor of a flat hierarchy. While this is freeing as it uncaps staff’s potential in the workplace, Victoria Weller, who oversees operations at ElevenLabs, is having to invest more in improving mechanisms and recognition to make up for the lack of titles.

“We don’t want there to be a ladder,” Weller says. “If you’re excellent, you can accelerate, you can move across different roles, and that doesn’t mean that you’re not growing your career.”

The move to remove titles at ElevenLabs has acted as a screening process for culture because it naturally rules out high-ego individuals who care more about how they are perceived in a process and therefore wouldn’t fit within the startup’s “impact over status” culture, Weller says.

“It is easier having low-ego individuals and trusting that they are motivated by achieving great results,” she says. At ElevenLabs, new ideas are taken at face value, rather than filtered through questions of who’s advancing them and why.

Performance reviews

With the pace of delivery so intense and roles more fluid, many AI startups are finding the traditional performance review system constraining. Self-driving startup Wayve has shunned old-school annual performance reviews to design a behavioral career framework with quarterly touchpoints, according to Emma Baillie, vice president of people. Quarterly check-ins can better fit the requirements of a pre-revenue startup, where goals can change rapidly and there’s not one clear steady path for staff to follow.

Laura Gonzalez, chief of staff at AI video platform Synthesia, is also moving toward a quarterly timeline that better reflects how the team works. “It’s not about a year or two years. We think much more short-term,” Gonzalez says.

In addition, Synthesia considers how to allow space for failure as part of the performance management process and in everyday work. A failure at Synthesia is considered “successful” when it produces valuable learnings at a low cost. One way it incentivises this learning process is by continuously launching new features – alongside weekly updates – which normalizes the process of iteration, makes progress visible, and reduces “all-or-nothing” pressure.

“When you’re constantly shipping and adjusting, pivots feel like normal course-correction,” Gonzalez says. “Achieving product-market fit is always messy and constant iteration is the only path to success.”

Tough conversations

Even when big-bet failure is tolerated, consistent underperformance can’t be. Whereas a decade ago the adage for top startups was often “hire slow, fire fast”, in the AI age we believe it’s more like “hire fast, fire fast”. While recruitment is moving at speed, it’s still important to trim weaker employees from the payroll before they fester.

“The definition of great remains the same, but there’s no tolerance for underperformance as was so common in the Covid era,” says Stan Massueras, who leads a large part of the go-to-market team in Europe for AI voice startup ElevenLabs.

Teams of individuals

As AI startups scale, one of the biggest challenges is collaboration under pressure. What happens when a team full of strong founder personalities face a crisis? A security vulnerability? A missed milestone? A strategic disagreement between people who are all used to being right? This is where many founder-dense teams can break down.

Effective approaches to this challenge separate founder energy from authority. One of our portfolio companies, Harvey, navigates this well. The co-founders carry the vision and the pace, while a professional executive team provides operational structure to absorb shocks without slowing the company down. It preserves the founders’ raw ambition without letting it become a bottleneck.

At 1X Technologies, which builds humanoid AI robots, the answer is candor: directness without blame, baked into how the team communicates. “What it does is create a more psychologically safe environment to actually speak your mind, to say the things that you need to say,” says Brian Bocchino, 1X’s head of talent.

AI job search tool Jack & Jill prioritizes their operating structure to get this right. While they could be in hiring hypergrowth, they are choosing to stay small with intention.

“Every person adds organizational coordination cost,” says Ed Steele, founding operator at Jack & Jill. “It adds up quickly.” The team works in person to enable staff to work quickly on whatever needs doing, even when roles can be fuzzy.

The scaling challenge

Different models are required at different stages of a business. As it scales, Synthesia has found that it takes more than just self-starting individual contributors to drive progress and has focused on matching its top talent in small, cross-functional “tiger teams” to ensure project success.

“You need to make sure everyone understands what we’re trying to do, and you need to create systems so information can flow, but then you need to allow people to move really fast,” says Synthesia’s Gonzalez.

Many startups successfully navigating expansion have one thing in common: they have structures in place to decide how decisions get made before a crisis, not during it. The same applies to culture broadly: each company has its own approach, but there’s a shared instinct to give exceptional people autonomy, hold them to a high bar, and build structures to let them do their best work without getting in the way.

The startups getting this right aren’t the ones with the most elaborate people processes. In our experience, it’s the ones who figured out culture early, named it and have had the discipline to hire and fire by it. In a market moving so fast, culture isn’t just a nice-to-have; we believe it’s a critical operating system. Get culture wrong, and the high-caliber talent you worked so hard to find may either leave or – worse – pull the startup in the wrong direction.