Over the last four weeks, we hosted two networking events, one in New York and one in Frankfurt. Two different locations, one underlying trend.
From my point of view, companies are following three radically different approaches to AI.
๐ญ. ๐ง๐ต๐ฒ ๐ฒ๐ ๐ฝ๐ฒ๐ป๐๐ถ๐๐ฒ ๐๐ฝ๐ฒ๐น๐น ๐ฐ๐ต๐ฒ๐ฐ๐ธ๐ฒ๐ฟ
Companies that rolled out Copilot with no underlying strategy. No defined use cases, no measurement, no change to the operating model. The result is an annual spend running into six or seven figures that, at best, polishes emails. The technology was deployed. The business wasn't transformed.
๐ฎ. ๐ง๐ต๐ฒ ๐ฐ๐ผ๐๐ ๐ฟ๐ฒ๐ฑ๐๐ฐ๐๐ถ๐ผ๐ป ๐๐ฟ๐ฎ๐ฝ
Companies treating AI purely as a headcount lever. The thesis sounds tidy on a slide: replace junior roles, compress teams, capture the savings. The reality, as voices inside OpenAI and Anthropic are now openly acknowledging, is messier. When you cut the junior pipeline, you also cut the path that produces tomorrow's senior judgment. The Wall Street Journal recently profiled engineers warning of "vibe slop": work shipped fast by AI that accumulates technical and operational debt and eventually catches up with the balance sheet. Cost reduction strategies also have a ceiling, usually a wall you don't see until you hit it.
๐ฏ. ๐ฅ๐ฒ๐๐ฒ๐ป๐๐ฒ ๐ฝ๐ฒ๐ฟ ๐ฒ๐บ๐ฝ๐น๐ผ๐๐ฒ๐ฒ
The third group, and the smallest, treats AI as a multiplier on the people they already have. They use AI to compress operational friction so talent spends more time on the work that actually compounds: client relationships, judgment, creative problem solving, and category creation. AI also lets their teams produce outcomes in faster cycles, outpacing the competition. Their headline metric isn't cost saved. It's revenue per employee, and it's climbing.
The first two camps are optimising for the past. The third is building for the next decade.
If you'd like to join the debate, DM me for details on our next events in NYC, Amsterdam, and Dรผsseldorf.