A decision, not a product category

Automating a job and hiring for it are not the same trade.

The two options get compared on monthly cost, which is the one dimension where they look comparable and the one that matters least. A subscription and a person differ in what happens when the work changes, when it goes wrong, and when you stop. Those are the terms worth comparing on.

The two branches

Set against each other on the things that actually differ.

Automate it

A tool, a subscription, and your own time to run it

What you pay: A monthly fee, usually per seat or per usage, that moves with how much you use it.
What it costs beyond that: Your time setting it up, and your attention every time it produces something that needs checking.
When the work changes: You reconfigure, or you find the tool does not bend that way and start again.
When you stop: You cancel. Whether your work comes with you depends on the export terms.

Hire for it

A person, contract or employed, who owns the outcome

What you pay: A rate or a salary, plus the employer costs that sit on top of it.
What it costs beyond that: Recruiting, onboarding, and the management time that does not stop after the first month.
When the work changes: You explain it once. Adapting is the thing a person is better at than a tool.
When you stop: Notice, process, and whatever only that person knew leaving with them.

There is no general answer, and anyone offering one is selling something. The useful version of this question is narrower: for this specific job, how often does it change, and how bad is it when it is wrong? Those two answers decide it more reliably than cost does.