‘Agency’ trumps modeling
VCs and TA leaders are reportedly prioritising 'agency'—initiative and independent judgement—over routine skills like modeling as AI handles repetitive tasks. (x.com)
Venture capital firms and recruiting teams are starting to prize “agency” — the ability to act without step-by-step instruction — as artificial intelligence takes over more routine work in screening, research, and analysis. (bloomberg.com) In hiring, “modeling” has long meant the structured tasks that junior staff used to prove they could do: build spreadsheets, sort résumés, run comparisons, and package findings for a manager. New artificial intelligence tools now do much of that work faster, and talent teams say the bottleneck has shifted from producing output to deciding what to do next. (forbes.com) Josh Bersin Company said in September 2025 that companies using artificial intelligence in talent acquisition were hiring 200% to 300% faster, and that one major United States resort operator increased scheduled interviews by 423% in 12 months after deploying conversational artificial intelligence. The same report said only 17% of applicants reached the interview stage in 2024 and 60% abandoned slow application portals. (prnewswire.com) Venture firms are making the same adjustment inside investment teams. Bloomberg reported on February 20, 2026 that firms are hiring people who can use artificial intelligence as part of the job, and that Felicis hired former Google product manager Ben Mathes this year as head of artificial intelligence to oversee internal tools. (bloomberg.com) That changes what employers test for. If software can search old memos, summarize calls, rank candidates, and draft first-pass analysis, the remaining work is choosing the right question, spotting what the system missed, and pushing a process forward without waiting for permission. (bloomberg.com) (forbes.com) OpenAI has been making the same labor-market argument in public. In September 2025, the company said businesses want workers who are fluent in artificial intelligence and announced plans for a jobs platform meant to connect employers with candidates who can use those tools. (openai.com) In an April 2026 policy paper, OpenAI said frontier systems had moved from helping with tasks that take minutes to tasks that take hours, and said future systems may be able to carry out projects that now take months. The paper also warned that jobs and whole industries could be disrupted as that shift spreads. (cdn.openai.com) Recruiters and investors are not saying technical skill no longer matters. Bloomberg reported that venture firms have spent the past few years hiring more technical investors, including people from OpenAI, and are now adding a second filter: whether candidates can use artificial intelligence well enough to improve the firm’s own work. (bloomberg.com) The result is a quieter rewrite of entry-level merit. When machines handle more of the repeatable tasks that once signaled competence, employers have fewer reasons to reward polish alone and more reasons to look for judgment, initiative, and follow-through. (prnewswire.com) (forbes.com)