AI hiring boom could become a surprise tailwind for software, Bernstein says

Artificial intelligence could turn into an unexpected positive for software companies if productivity gains prompt businesses to add employees rather than reduce headcount, Bernstein analysts argued, challenging one of the sector's main bearish narratives.

Worries that AI will displace white-collar staff have weighed on the outlook for application software vendors, which have historically charged customers through seat-based licenses. A smaller workforce could mean fewer paid software seats.

Bernstein contends the opposite result may be emerging. Research based on Ramp data showed that companies with heavy AI adoption were recruiting more aggressively than businesses with low or no adoption. On average, heavy adopters recorded headcount growth more than 10% above control companies and their lower-adoption counterparts.

That could help software vendors maintain seat-based pricing alongside newer models tied to AI token consumption, infrastructure usage, or return on investment.

Enterprise clients also continue to prefer seats partly because costs are predictable, Bernstein said. If AI-driven productivity leads companies to expand their workforces, seat-based licensing may become less of a concern for software firms.

The hiring increases were not limited to engineers and salespeople. Administrative, customer service, and entry-level roles also expanded among heavy AI adopters, contradicting expectations that AI adoption would disproportionately eliminate junior and back-office jobs.

Bernstein compared the trend with earlier productivity booms triggered by PCs and the internet, when companies largely reinvested efficiency gains into growth rather than sharply widening profit margins. White-collar employment rose during the PC era despite significant productivity improvements.

There are caveats. Finance, scientist, and operations positions showed weaker or no statistically significant positive growth, suggesting that software aimed at those functions may receive less support.

Bernstein made no changes to its software models, price targets, or recommendations based on the analysis.