How much does a virtual assistant cost in Latin America?

Pricing in this category is deliberately hard to compare. Some providers quote an hourly rate, some a monthly retainer, some a per-task credit, and almost nobody quotes the number you actually need — what the work costs you all-in, once management time and turnover are counted. This is our attempt at the honest version.

What the market actually charges

Three bands cover almost everything on offer, and the difference between them is not mostly geography. It is how much judgement the person is expected to exercise without being asked.

Full-time, the nearshore band works out to roughly $1,300–$2,800 a month. Specialised profiles — bookkeeping, paid media operations, technical support — sit at the top of it or just above. Executive assistant work with calendar ownership and stakeholder contact runs $15–$30 an hour.

Why the range is so wide

Four things move a quote inside that band, roughly in order of how much they matter:

1. 01Judgement. Executing a documented process is one price. Deciding what the process should be is another. Most of the spread is here.
2. 02Specialisation. General admin is the floor. Anything with a certification, a regulated output, or a tool that takes months to learn is above it.
3. 03Language. Genuinely bilingual delivery — writing to your customers in English, not just understanding your instructions — carries a premium of roughly 15 to 25 percent.
4. 04Dedication. A person assigned to you costs more per hour than a shared pool, and is the only arrangement where knowledge of your business accumulates.

The comparison most quotes leave out

Comparing an assistant's hourly rate to an employee's hourly wage is the most common mistake in this decision, and it is not close. In the US private sector, benefits account for 30.1 percent of what employers actually spend on compensation: wages average $32.60 an hour and benefits another $14.01, per the Bureau of Labor Statistics.

Add payroll taxes at 7.65 percent, workers' compensation, unemployment insurance, and the desk, laptop and software the person needs, and the working rule is that an employee costs 1.25 to 1.40 times their base salary. A $50,000 role is a $62,500–$70,000 line item.

The costs that do not appear on the invoice

Two of them are large enough to change which option is cheapest.

The first is your own time. Every assistant needs onboarding, and someone has to answer questions during it. Budget four to six hours in week one and an hour a week after that. If you are the bottleneck the arrangement is meant to relieve, that cost is real and it is yours.

The second is turnover. Replacing an assistant means paying the onboarding cost again and losing whatever was never written down. This is where the cheapest hourly rate usually stops being cheapest: marketplace work has high churn by design, so you pay the setup cost repeatedly and nothing compounds.

How to tell whether it is worth it

The arithmetic is simple enough to do before you talk to anyone. Take the hours a week you would hand over. Multiply by what an hour of your own time is worth — for an owner, revenue divided by hours worked is a fair proxy. Compare that to the monthly rate.

Handing over ten hours a week of work you value at $60 an hour is $2,400 a month of your time against $1,300–$2,800 of cost. That is roughly break-even on paper, and clearly positive in practice, because the hours you get back are the ones you cannot delegate and the hours you hand over are the ones anyone can do.

If the sum does not clear on paper, the honest answer is that you do not need an assistant yet. You need to automate the task or stop doing it.

How to structure a trial before committing

You do not have to decide the full arrangement up front, and you should not. A month designed as a test tells you more than a quarter designed as a commitment.

1. 01Pick two processes, not ten. Something recurring and low-risk — inbox triage and invoicing are the usual pair.
2. 02Write each one down before anyone starts. Half a page is enough. If you cannot, that is the finding: the process does not exist yet, it lives in someone's habits.
3. 03Set the number you will judge it on. Hours returned, response time, backlog cleared. One number, agreed in writing.
4. 04Give it four weeks. Two is not enough to get past onboarding; eight is long enough that you stop measuring.
5. 05Decide at the end on the number, not on how the month felt. Months feel busy either way.

The most common reason a trial fails has nothing to do with the assistant. It is that nobody on the client side was available to answer questions in week one, so the assistant spent the month guessing. Budget four to six hours of your own time that week, or postpone the start until you have them.

What "pre-trained" should mean

The phrase is used loosely across the category, so it is worth pinning down. It should mean the person already works fluently in the tools most businesses run on — a shared inbox and calendar, a spreadsheet, a CRM, an invoicing tool, a project board — so onboarding is about your processes and not about software basics.

It should not be taken to mean trained on your business, which nobody can be in advance. A fair test when comparing providers is to ask which specific tools are covered and what happens when you use one that is not on the list.

How we price it

Delfín Force scopes monthly rather than hourly. Not for margin reasons: an hourly count makes both sides optimise the wrong thing, because we get paid for time spent and you start auditing minutes instead of outcomes. Scope is agreed before starting, reviewed as the work changes, and there is no long lock-in.

Onboarding takes about 48 hours from the moment scope is agreed, because assistants arrive already trained on the tools most businesses run on. What remains is learning your processes, not the basics.