What This Is
A 2-person AI team is expected to move 5 departments — the lead driving AI deployment at an enterprise recently proposed establishing an "AI Strategic Development Committee" to the company. What he's solving isn't a technical problem; it's an organizational coordination problem.
The AI business unit has only two people, and projects span five departments: clinical, marketing, training, after-sales, and development. In the article he describes concrete situations: the clinical agent needs professional content review, marketing must organize trials, training and after-sales both need to coordinate. Every department is busy, every department has something more urgent. What consumes the project lead's day isn't writing code — it's chasing requirements, materials, and confirmations. Technical overtime solves technical problems; other people's priorities can't be reshuffled by staying late.
His solution isn't adding headcount — it's getting the chairman to back the initiative and appointing a secretary-general to drive execution, handling cross-departmental resource competition, scope changes, and budget execution — these "concrete conflicts." Day-to-day technical issues stay with the execution team; only matters beyond the lead's authority go to the committee.
This article isn't about technology. It's about a question that's repeatedly avoided: the real bottleneck in AI deployment may not be that models aren't good enough — it's that organizations aren't ready.
Industry View
One view: forming a committee is a "luxury" for large companies; small and medium businesses can get by with project weekly meetings and enterprise WeChat groups. But the conflicts described in the article come precisely from a small team — 2 people against 5 departments — with naturally weak bargaining power. We note that without formal authorization, the AI lead degenerates into a "human reminder system" — however strong their technical ability, they can't deliver.
Another question: doesn't this "add a layer of waiting"? The author himself flags this risk, so the design keeps day-to-day technical issues with the execution team and only escalates resource conflicts to the committee. Whether it actually works depends on how tightly the secretary-general follows up — the hardest-to-supervise link in any organizational design.
Another point the author repeatedly warns about: don't slap the "AI-native organization" label on yourself. They consider themselves still in the launch phase; daily workflows and collaboration methods need gradual adjustment. This kind of restraint is rare in the AI hype cycle.
Impact on Regular People
For enterprise IT: if your company is also pushing AI projects, technology selection and development scheduling may not be the hardest part — getting business departments to match your pace is. This article is worth sharing with your boss, to help them understand that AI projects need formal authorization, not just goodwill.
For individual careers: colleagues who participate in AI projects — organizing materials, reviewing content, running training — whether these contributions are seen and rewarded determines whether they're willing to keep spending time on the next round. This is the hidden cost of project sustainability.
For the consumer market: short-term impact is minimal. But the "using it" phase after an AI product launches — whether feedback loops back and the system keeps iterating — often determines whether it's still on your phone three months later.