Most companies built before the generative AI era are now operating on borrowed assumptions, according to Gigi Levy-Weiss, co-founding partner at venture firm NFX. Speaking at EY Israel's JOURNEY 2026 conference at the Hilton Tel Aviv, he argued that the overwhelming majority of founders would design their companies differently if they started today - and that the gap between the old design and the new reality is where competitive risk now lives.
Why the Org Chart Became the Problem
Levy-Weiss's central claim is structural rather than technological. Traditional org charts, he said, were built around specialization: each department held narrow expertise, and work moved between them in sequence. That model made sense when knowledge was scarce and unevenly distributed. It makes less sense now, he argued, because every handoff between departments - from product to marketing to engineering - loses information and time. "Your org chart basically encodes the standards of the past," he told the audience.
His proposed alternative is the "pod": a small team holding full responsibility for a problem from the moment a customer need is identified to the moment a solution reaches that customer. What makes pods viable now, in his view, is a shift in what individual employees actually do. People, he said, have become orchestrators of automated systems rather than executors of narrow tasks - which allows a handful of generalists to cover ground that once required an entire department.
What NFX Calls the "Second Founding"
NFX frames this shift as a deliberate, non-incremental process it calls the second founding. Levy-Weiss was explicit that using tools like ChatGPT or Claude to write code or draft presentations does not qualify. Those applications, he said, are "table stakes" - useful, but not transformative. A genuinely AI-native company, by contrast, changes the actual path from identifying a customer need to delivering a solution. If that path looks the same as it did three years ago, the company has not made the shift, regardless of how many AI tools its employees use.
This reordering also changes how problems get approached inside a company. Instead of starting with "which people do we assign to this," the first question becomes whether the problem can be automated with AI or software agents. Only afterward does the question of human staffing arise - focused on who is needed to run, supervise, and govern those automated systems rather than perform the underlying task.
Speed, Experimentation, and the Human Cost
Levy-Weiss argued that the only meaningful measure of speed is how quickly a company can put a working solution into a customer's hands. Faster cycles enable more experiments, and more experiments mean more chances to find what works - a logic that rewards iteration over careful, department-by-department planning.
He did not soften the implications for staffing. Every role, he said, needs to become a builder role, including functions like finance and marketing that have traditionally sat apart from product development. He acknowledged plainly that many people currently employed by companies may not be suited to the roles those companies will need in the future, describing this as a moral question he was raising but deliberately not resolving.
Notably, he extended the critique to founders and executives themselves, including his own ambitions as an investor, arguing that past achievements were shaped by constraints that no longer apply. His closing instruction to the room was direct: build a company capable of surpassing its own founders, because if they do not do it, a competitor eventually will.