Ten people used to mean receptionist, bookkeeper, marketer, support, ops — each role a slice of the day. one founder running a lean AI-assisted company can cover more of those slices than was realistic five years ago.
That does not mean a ten-person company collapses into a laptop. It means the shape of “small team” changed.
What one person plus AI can actually run
Drafting and scheduling content. First-pass invoicing and expense sorting. FAQ-style support with escalation rules. Simple CRM hygiene. Research packs before sales calls. Prototypes that used to need a freelancer for every mock.
The bottleneck moves from “who will type this” to “who will decide and take responsibility.”
What still needs humans — or more than one
Relationships that require trust over time. Regulated work. On-call coverage. Domain expertise you cannot prompt into existence. Conflicts, negotiations, and moments when a wrong answer costs a client.
Burnout is also a headcount problem: one person owning every role has no backup.
A realistic framing
The hidden line item is often counting AI token costs alongside headcount.
Eventually finance asks for checking whether an AI investment really pays off.
AI can replace tasks that once justified hiring. It rarely replaces the need for specialised judgement at scale. A solo operator with strong tools can punch above their weight — and still hit a wall when the company must grow beyond what one calendar can hold.
Brak komentarzy:
Prześlij komentarz