The operator skill stack of the last fifteen years was well-understood. You knew how to design a sales motion, plan headcount, manage a quarterly roadmap, run a board meeting, and translate ARR into a financing strategy. Those skills still matter. They are no longer the skills that decide who builds the next generation of important companies.
The new stack does not replace the old one — it sits on top of it. The operators who add the new layer fastest will pull away from the ones who do not.
What the SaaS-era operator was good at
The classical operating skill set was built around scaling a stable product through predictable channels. Growth came from headcount: more salespeople, more support agents, more engineers shipping a long roadmap. The skills that mattered were the skills of running a larger and larger version of the same machine.
That worked because the underlying product, market, and channel did not shift while the company was scaling. The operator's job was to build the org that could meet the demand the product was generating.
What changes in the AI era
The product is no longer stable. The model improves under you, often weekly. Customer expectations shift faster than the roadmap can be planned. Workflows that used to require a team are absorbed into the system. Headcount is no longer the lever for scaling output, because the marginal unit of work is increasingly performed by the product itself.
The operator's job is no longer to scale the machine. It is to keep redesigning the machine while it is running.
The new stack
- Judgment under model uncertainty. The model you ship on today is not the model you will ship on next quarter. Operators who plan as if the ground will move — and design products that get better when it does — out-perform operators who plan as if the current capability set is the ceiling.
- Taste in selecting what to automate vs. keep human. Almost every workflow can now be partly automated. The question is which parts. The operator who knows where the customer values human contact, where the system can absorb the work invisibly, and where automation actively subtracts trust is making decisions that will matter for a decade. This is taste, not technique.
- Compounding a small team into outsized output. The companies that will compound for twenty years are the ones whose output per person continues to rise. That requires operators who know how to keep teams small, decisions tight, and tools sharp. The opposite instinct — adding people to solve problems — is the single most expensive habit of the previous era.
- AI-fluent product instinct. Knowing what AI can and cannot do well today is a craft skill. Operators who have used the tools daily have an edge over operators who have read about them. The gap shows up as faster product decisions, better scoping of new features, and fewer dead-end builds.
- Capital efficiency as the default. The cheapest operator is the one who has done it before with less. Money is now a worse competitive advantage than it has been in any other tech cycle — because the things money used to buy (engineering throughput, distribution reach, content output) are being collapsed in cost. Operators who have built inside that constraint will be the ones who run the most efficient AI-era companies.
- Comfort working alongside AI agents as part of the team. The org chart of an AI-native company has humans and agents on it, often interchangeably. Operators who can design workflows where both contribute usefully — and who can hire humans for the parts only humans can do — will move faster than operators who treat AI as a tool rather than as labor.
- Ruthless focus on retention. Acquisition is getting cheaper. Retention is getting more important. The operator who designs the company around making customers measurably more capable each month is the operator who builds the durable business. Everyone else is competing on a deflating channel.
What is becoming less valuable
A few skills that previously sat near the top of the operating stack are quietly losing weight.
- Headcount-driven scaling. Adding bodies to solve a throughput problem is no longer the cheapest move. Often it is no longer even a move.
- Long roadmaps with quarterly increments. The companies that ship best now run on smaller, faster cycles than the SaaS era taught us. Quarterly roadmaps treat the product as more stable than it actually is.
- Sales-led growth as the default motion. Sales-led growth is still the right answer for some categories. It is no longer the right answer for most. Operators who default to it underestimate how much value is now being created in product-led and distribution-led models.
What we look for in a founder
Inside our venture studio and across our investments, the founders we want to back are the ones who have internalized the new stack. They are usually ex-operators who already have an opinion about a market we both care about, who can take a concept from validation through launch in months not years, and who would rather ship a tighter team and a sharper product than scale on someone else's capital.
The new operator stack is not credentialed. There is no MBA program that teaches it. It is learned by building, mostly inside companies that have already been pushed through the AI rewrite. We are looking for the people who have been doing it.
Why this matters beyond founders
The skills above are not just for CEOs. They are the skills that define a senior IC, a leader of a small team, a partner inside a professional-services firm, a founder-in-residence, an operating partner inside a holding company. Anyone whose job is to translate intent into outcome under conditions that change weekly is operating against the new stack — whether they have named it or not.
The economy is rebuilding around people who can. Worth investing in.