Will AI-native firms require fewer workers?

Companies built around artificial intelligence employ roughly 25% fewer workers than their peers while achieving comparable or higher valuations, according to a new Harvard Business School working paper

The paper, titled "AI-Native Firms," was co-authored by Hyunjin Kim of INSEAD and Rembrand Koning of Harvard Business School. The researchers studied nearly 2,900 Y Combinator startups from 2020 to 2024, then extended their analysis to tens of thousands of venture-backed companies tracked by PitchBook, linking both datasets to workforce records from Revelio Labs.

Kim and Koning found that AI-tagged startups are about 25% smaller than non-AI startups in the Y Combinator sample and 12% smaller in the broader PitchBook sample, after controlling for industry and cohort. In raw terms, the average AI startup in the Y Combinator sample employed approximately 13 workers, compared with 42 for non-AI firms.

Despite the smaller headcounts, these firms raise comparable or greater funding and achieve similar or higher valuations than non-AI competitors. In the PitchBook sample, AI-tagged firms achieved 76% higher valuations per employee than their non-AI counterparts. Kim and Koning write that this pattern is consistent with AI "complementing technical expertise at the firm level."

Product over process

The study distinguishes between two ways AI can change a firm: a "process channel," in which staff use tools such as ChatGPT or Claude to do existing jobs faster, and a "product channel," in which AI capabilities are built directly into what the company sells. The authors find that only the second channel explains the organizational differences they observe.

Job postings at AI-native firms name specific tools such as Cursor or GitHub Copilot at roughly 2.6 times the rate of non-AI firms, according to the paper. However, this measure does not predict smaller headcounts once the researchers control for other variables in their models.

By contrast, companies classified as embedding AI directly into their products employ roughly 10 fewer staff than otherwise comparable peers. Kim and Koning write that "if this pattern holds, then building, importing, and orchestrating model capabilities may matter as much for competitive advantage as building human ones." The authors cite FazeShift, an accounts-receivable automation company, where a team of ten performs invoicing and collections work that a conventional vendor would typically require a much larger staff to deliver.

Flatter hierarchies, Fewer entry-level roles

AI-native firms also employ significantly fewer managers and have hierarchies roughly half a seniority level flatter than non-AI companies. Kim and Koning report that AI startups have about 15% fewer managers as a share of total employment.

Entry-level employment shares are also lower at AI-native firms, down roughly four percentage points, while senior-employee shares are correspondingly higher. The authors note this runs counter to research suggesting AI tools tend to benefit less experienced workers most at the individual task level.

This effect is strongest in "services-delivery" industries, including healthcare and media, where AI-native firms operate at roughly 30% the headcount of non-AI peers in the same industry and cohort. The authors attribute this to service-based work — such as tutoring or customer support — that historically required large staffs now being performed inside AI-powered products.

Kim and Koning caution that their findings are correlational rather than causal, and that shrinking headcounts at individual firms do not necessarily translate into lower aggregate employment across the economy.

“AI may not simply make existing organizations more efficient—it may change what organizations look like and do. If firms can increasingly import capabilities from foundation models rather than build them through people, they may no longer be

constrained by human attention and processing and the managerial problem shifts from accumulating internal capacity to building and integrating external capabilities into products and production,” the authors wrote in their report.

“The startups we study offer an early window into this shift.”

Three in 10 (30%) HR leaders in the U.S. say their talent acquisition strategy is shifting towards hiring fewer entry‑level workers in favour of mid‑level employees using AI to complete what were previously junior tasks, according to a previous report.

Here are some companies that have cut jobs amid AI adoption: