# Should You Hire a VA or an AI Employee for Your Business

*Question — 2026-09-18 — by Mahmoud Zalt*

A five-question test for deciding between a virtual assistant and an AI Employee, with a worked example and the honest case for hiring both.

**Short answer.** **Decide by looking at the work, not the price tag.** If the jobs you want off your plate change shape every time, hire a VA, because someone has to make the calls. If they repeat with a known shape, hire an AI Employee, because that work punishes a person and suits software. Run the five-question test below and the answer usually falls out in under ten minutes.

This decision tends to stall for a reason that has nothing to do with either option. You are not really choosing between two kinds of help. You are trying to describe your own week clearly enough to hand a piece of it away.

So the test below is not really about software. It is a set of questions about the work itself. Answer them honestly and the spending decision stops being a matter of taste.

For context on the other side of the comparison: an AI Employee on **Sistava** is hired into a named role. You write a short brief, connect the tools it needs, and give it standing duties. It then works in your inbox, CRM and documents on a schedule, rather than waiting for you to open a chat window and think of something to ask.

## At a Glance

- **5** Questions in the test below
- **10 min** Time it takes to get an answer
- **2 weeks** How long to run a fair trial before judging
- **Most** Small teams who end up hiring both, in that order or the reverse

## How do I decide between a VA and an AI Employee?

Take the single job that annoys you most this week and put it through five questions. Each one has a clear side. If four or five answers point the same way, that is your hire. If they split three to two, you are looking at two jobs pretending to be one, and you should separate them before spending anything.

This works better than a pricing comparison because the cost of a bad delegation is never the monthly fee. It is the three months you spend supervising work that was handed to the wrong kind of helper.

### The five-question test

1. **Could you write this down for a new starter in one page?** — Yes points to an AI Employee. No, or not without a long conversation, points to a person.
2. **Does the right answer depend on knowing a specific human?** — If the work needs someone who understands a client's mood or history, that is a person's job.
3. **Does it happen more than twenty times a month?** — High repetition is where an AI Employee earns its keep and where people quietly burn out.
4. **Does anything happen off the screen?** — Posting, signing, collecting, attending, phoning a supplier who never replies to email. All human.
5. **Would a mistake cost money or trust?** — If yes, you need a review step no matter who does it. That is not a tiebreaker, it is a condition.

## When should you hire the VA?

Hire the person when the work is still being invented. If you cannot describe the process because there is no process, no amount of software will help, and you will spend your evenings writing briefs for a job you have not figured out yet.

Hire the person, too, when the work involves other people caring who is on the other end. Clients who are stressed, suppliers you need a favour from, a candidate you are trying to keep warm. Those are relationships, and relationships need someone to own them.

There is a third case people miss. Hire the person when you need someone to build the process so that it can be handed over later. A good assistant spends their first month turning your chaos into a document, and that document is worth more than the month costs.

- The work has never been written down and changes every time
- Someone needs to make judgment calls you would stand behind
- A client or supplier relationship is at stake
- Anything physical, in person, or on an unpredictable phone call
- You want a second brain in the business rather than a second pair of hands

The reason the test works is that it stops you comparing a person to software in the abstract. It makes you look at one specific pile of work and ask who it suits. Here is what that looked like for someone who was about a week away from posting a job advert.

## A worked example: Yusuf, a commercial property broker

Yusuf brokers small commercial leases in Chicago on his own. He had budgeted around 1,800 a month for a part-time assistant at twenty hours a week, and he had a shortlist of three candidates ready to interview.

Running the test changed the shape of the hire rather than cancelling it. Two jobs failed every question for software: showings, where he needs someone physically at a building, and the awkward middle of a negotiation where a tenant is wobbling and needs a phone call from a human being.

Three jobs passed every question for an AI Employee: writing up listings from his site notes, replying to the first round of enquiries within minutes rather than the next morning, and keeping his pipeline notes current. Together those were around 60 enquiries and 9 write-ups a month, all of them shaped the same way every time.

He hired the assistant at eight hours a week instead of twenty, and briefed an AI Employee on the other three jobs. The assistant now does showings and the difficult calls, which is the part of the job that actually wins deals. The enquiry that lands at 9pm gets a real reply at 9pm, which is the part that used to lose them.

## What the AI Employee will not handle in this decision

It will not tell you which jobs to hand over. That judgment is yours and it does not get easier with a better tool. Plenty of people hire one, hand it their vaguest task, and conclude the whole category does not work.

It also will not carry the consequences. If a listing goes out with a wrong number, someone has to notice and fix it, and that someone is a person on your team. You get a worker that never tires, not a worker that takes responsibility.

**The mistake that wastes the most money.** Handing over the job you understand least. It feels like the biggest win, because it is the one you hate. It is also the one you cannot brief properly, so the output comes back wrong and you conclude the tool is useless. Start with the job you could explain to a new starter in five minutes.

## Comparison

| Dimension | Traditional | With Sista |
|---|---|---|
| The process | Does not exist yet and needs building | Exists and could be written on one page |
| The frequency | A handful of times a month, each one different | Dozens of times a month, each one alike |
| The stakes | Someone's trust in your business is on the line | Speed and consistency matter more than nuance |
| The timing | Working hours are fine | Nights, weekends, and your holidays all need cover |
| The failure mode | Nobody is thinking about the problem | The work is simply not getting done fast enough |
| What you are buying | Hours of attention and judgment | A standing role from $25 a month |

If both columns describe parts of your week, that is a normal result rather than a failure of the test. It means you have two piles, and the useful move is to size each one before you spend. Most people find the repeatable pile is bigger than they assumed and the judgment pile is smaller but far more valuable per hour.

One thing worth saying plainly before the questions: if you already have a virtual assistant you like, none of this is an argument for letting them go. The teams that get the most out of both usually keep the assistant and change what fills their week, so the hours you pay for go to the work only a person can do.

## Common questions

## FAQ

### If I can only afford one this quarter, which comes first?

Count how many of the jobs on your list you could explain to a new starter in five minutes. If it is most of them, start with the AI Employee, because that work hands over cleanly and immediately. If it is only one or two, hire the person, because the rest of your list needs someone to make decisions before it can be handed to anything.

### Can I hire an AI Employee first and add a VA later?

Yes, and that order works well for solo founders. The AI Employee clears the repeating work, which usually reveals what is actually left, and what is left is a much clearer job description for a person. You end up hiring a better-defined role than you would have advertised six months earlier.

### How long before I know if it was the right call?

Two weeks of real work, not two days of testing. The first few days are always rough because your brief is incomplete, and judging then tells you nothing. By the end of the second week the error rate should be visibly falling, and if it is flat, the job was probably the wrong shape.

### Will an AI Employee make my existing assistant's job insecure?

That depends entirely on what you fill their week with afterwards. If you take the tedious half of their work and give them the client conversations and process building instead, most assistants are relieved. Be direct about it early, because the uncertainty is worse than the change.

### What if the job is half repeatable and half judgment?

Split it. Almost every job that feels ambiguous is actually two jobs stapled together, usually a drafting step and a decision step. Let the AI Employee draft and let a person approve, which is also the safest starting arrangement for anything customer-facing.

If your list is mostly customer conversations rather than admin, there is a companion piece worth reading before you spend anything. It goes through the specific situations where hiring a person is clearly the better use of the same money, written without any attempt to talk you out of it. Read that one if the five-question test came back leaning human.

Run the five questions on the single job that annoyed you most this week. Not the biggest one, the most annoying one, because that is the one you will actually follow through on. Whichever way it points, you will have spent ten minutes and come out with a clearer description of your own business, which is useful whether or not you hire anything at all.

**Tags:** hire-a-va, ai-employee, delegation, small-business, decision-guide