An AI employee vs hiring a person
This one deserves a straight answer, because it is rarely a choice between a person and a piece of software. It is a choice about which half of a role you are actually trying to fill.
At a glance
The comparison is about work, not headcount
Most work inside a company is coordination, not judgement. Chasing, reconciling, triaging, scheduling, keeping a record current, assembling the same report by hand every Monday — it fills the week of people who were hired for something else, and none of it is why they took the job.
So the useful question is not “person or AI employee”. It is: of the work in this role, how much is coordination and how much is judgement? Hand over the first half and you are not replacing anyone; you are giving the second half back the hours it needs.
Cost is the least interesting difference
Pricing is per AI employee on shift — $290 a month on Starter, $240 on Company — and humans are unlimited on every plan, so nothing about growing the team changes the bill. Retire an employee and billing stops at the next cycle. There is a 14-day trial with no credit card, and you can cancel any time.
That is easy to hold up against a salary, and it is the wrong comparison. A salary buys judgement, relationships and someone who notices when the situation has changed. A subscription buys hours of coordination that never gets tired at 18:00. The number worth comparing is neither of those: it is what the work costs you when it is late.
Ramp, hours and the parts that do not tire
A new hire is an investment that pays back slowly and then compounds: they learn the business, the customers, the reasons behind the exceptions. There is no shortcut, and no reason to want one.
An AI employee arrives the other way round. It comes with a working memory of your business — your documents and connected systems — so there is no onboarding period, and the median time to a first employee on shift is 11 minutes. It is on shift continuously, in every timezone, with no rota, which matters most for the work that is only annoying because it happens at the wrong hour: the overnight reconciliation, the ticket that arrives in the middle of the night, the report someone assembles before anyone else is awake.
How to split a role
Describe the role in plain language — a job description is enough. Scope its systems: read, write or send, granted per system, revocable instantly. Then name the decisions that wait for a person, which is the part worth taking seriously; by default, anything with a consequence you would not want reversed.
You can run it in draft-only mode to begin with, where nothing leaves the building until someone presses approve, and tighten or loosen from there once you have watched a week of runs.
Where hiring a person wins
Judgement under ambiguity, and being accountable for it. When the situation is genuinely new, when the rule and the right answer disagree, when someone has to decide what the company is willing to live with — that is a person’s work, and Dutta is built to escalate to them rather than to resolve it quietly.
Relationships. A supplier who takes the call, a customer who trusts the person on the other end, a colleague who knows the history of a decision that was never written down. None of that transfers.
Decisions about people. Our terms prohibit using Dutta for automated decisions about people that require human judgement — hiring outcomes, credit, medical and legal decisions among them. A recruiting employee screens against your criteria with its reasoning shown, schedules, and keeps candidates answered; every advance, reject and offer is a person’s call.
And the case people forget: if nobody can write down what the process actually is, the process is not ready to be delegated — to a person or to anyone else. Hire the person, let them build it, then hand over the half that has become routine.
Questions
Do we have to let someone go for this to be worth it?
No. Humans are unlimited on every plan and pricing is per AI employee on shift, so this is not a swap. The comparison we would rather you make is against the work that is currently arriving late, or being done after hours by whoever cares most.Can an AI employee decide who we hire?
No, and our terms prohibit it. It screens candidates against criteria you set with its reasoning shown, schedules across panel calendars, keeps candidates answered and keeps your applicant tracking system current. Every advance, reject and offer is approved by a person.What happens when it gets something wrong?
AI output can be wrong, which is why consequential output is reviewed by a person and why you name the checkpoints. Every run is logged with its inputs, sources, reasoning trail and output, and any run can be replayed — so the question “why did it do that?” has an answer you can read rather than reconstruct.Decide it with your own process
The fastest way past a comparison page is a real one. Tell us the process that eats your team’s week and we will say plainly whether an AI employee is the right answer for it — a human replies, usually within one business day.
Talk to us · See what each role handles on shift · Pricing
Other comparisons
A general chat assistant · No-code automation · Doing nothing