Automation is strongest at tasks with a clear, repeatable pattern — logging enquiries, sending confirmations, answering the same five questions. It's not well-suited to judgment calls or relationship management.
A junior hire costs more than salary alone — training time, management overhead, and turnover risk. A well-scoped automation has an upfront build cost and a predictable monthly running fee, with no turnover risk.
Anything requiring genuine judgment, negotiation, or building an actual relationship with a client isn't a good automation candidate yet.
Automate the repetitive front-end — enquiry capture, FAQ handling, logging — and hire, or keep doing it yourself, for the judgment-heavy work that frees up.
It doesn't really stop making sense — the tasks it's good for exist at every size. What changes is whether you have enough volume to justify the upfront build cost.
That's usually the right answer — automation handles the repetitive load so a hire can focus on the parts of the job that actually need a person.
Automating something that genuinely needs human judgment and creating a bad customer experience, or hiring for purely repetitive work a well-built system would have handled for a fraction of the cost.