Virtual assistant agency services: models and operating fit

What to take away

Scores on this page run 1 to 5 and reflect what each service publishes about its own model. They are not a measure of how good any assistant is.

How these services were judged

Ranking by reputation is easy and useless. The four axes decide whether a service can be staffed at all.

Two services at opposite ends of any one axis should not sit with the same person in the same block of time. That rule prevents most quality problems in this trade.

Each provider is scored 1 to 5 on four items: pricing, named continuity, documented backup and named specialties. On pricing, the score reflects how openly the service sets out its model and what a quote depends on, not the amount it charges. On continuity, backup and specialties, a 5 means the arrangement is named and documented in the service's own material, and a 1 or 2 means the buyer carries more of that risk. The overall score is the unweighted average of the four items, and it is not a performance test.

The named services, scored and compared

Scores run 1 to 5. Overall is the unweighted average.

Time etc (timeetc.com) scores 3.0 overall, strong on price transparency (5) and limited on named continuity (2). It is a UK-founded, US-operating assistant service built on a subscription model: clients buy hours per month and use them across admin, inbox, research and scheduling. It suits owners who want a generalist on tap without hiring.

Pricing runs on monthly hour bundles rather than per task or per request, and that structure is what earns the 5 for price transparency.

For load, treat Time etc as queueable generalist work. Urgent same-hour coverage needs a stated window.

Prialto (prialto.com) rates 3.8 overall, strong on staffing continuity (5) and limited on price transparency (2). It runs a managed model with a named assistant backed by a trained production team, aimed at executives and small firms. It suits buyers who need continuity and a documented backup.

Pricing is quoted after a scoping call, and the quote depends on hours, assistant seniority and the size of the production team.

For load, Prialto fits executive support that can queue. The production team reduces single-person risk.

Athena (athena.com) rates 4.0 overall, strong on dedicated coverage (5). It pairs each client with a dedicated executive assistant and charges by the month. It suits founders with a heavy calendar and travel load, and its pricing is set by plan and by the assistant's hours rather than by task.

For load, Athena fits calendar and travel work. Task bursts need a separate plan.

Belay (belaysolutions.com) rates 3.8 overall, strong on multi-line vendor fit and limited on price transparency (2). It places US-based virtual assistants, bookkeepers and web specialists, and is one of the few that publishes a bookkeeping line alongside admin. It suits firms that want one vendor across two service families.

Pricing is quoted after a matching call, and the quote depends on the role, hours and service family.

For load, admin plus bookkeeping needs explicit access design.

Fancy Hands (fancyhands.com) rates 2.8 overall, strong on price transparency (5) and limited on dedicated continuity (1). It sells task-based requests rather than a named assistant: you submit jobs and they are distributed. It suits short, well-specified tasks, and its pricing is built on monthly request plans, so requests are counted rather than hours.

For load, Fancy Hands fits queued tasks. Context-heavy work will struggle.

Zirtual (zirtual.com) rates 3.8 overall, strong on backup (5) and plan clarity (4) and limited on specialist depth (2). It sells monthly plans with a dedicated assistant plus a backup. It suits small businesses wanting a named person at a lower commitment than a full-time hire.

Pricing is built on plan tiers, so the commitment is fixed before the first task is assigned.

For load, Zirtual fits ongoing admin with a named person and a backup.

Best for continuity: Prialto or Athena. Best for task bursts: Fancy Hands. Best for published pricing on generalist work: Time etc. Best for one vendor across admin and bookkeeping: Belay. Best for a named person plus a backup on a small plan: Zirtual.

The interruptibility rule, stated plainly

An assistant covering an inbox is not doing focused work. They are available. An available hour produces less than a focused hour, and pricing both the same means one of them loses money.

Sell availability as availability, with a stated response window and a volume cap. Or sell output as output, with a turnaround rather than a response time. Selling one and delivering the other is why clients say an assistant is slow: they are not slow, they are interrupted.

Where you must mix, block the day. Availability in defined windows, focused work outside them, and a written rule for urgent requests that land mid-block.

Access is a service design question

Bookkeeping support and systems work need access that, misused or leaked, damages the client badly. That is a reason to design the service around the access, not to refuse the work.

Three decisions belong in the written service definition: who holds the credential, what the holder can do with it, and how it is withdrawn when someone leaves. Put them in the scope so the client sees them.

Where a line touches a client's books, the obligation sharpens. The IRS guidance on which business records to keep sets out what must stay retained and retrievable. A bookkeeping line that leaves the client unable to produce those records has failed even if every entry was correct.

FTC compliance and line retirement rules

Your service description is an advertisement, and the FTC's advertising guidance for small businesses applies that to comparisons and implied claims, including claims you do not state outright.

Describe the work, the coverage and the review step. Those are yours to promise and yours to prove.

Retire a line when one client uses it, only your most senior person can deliver it, it produces most of your exceptions, or its cost has never been separated from everything else. Stop selling it, agree an end date, document the handover, then remove it from the site.

What changes as the bench grows

At one person the menu is whatever you can do. At four it is whatever you can document. At ten it is whatever survives someone leaving.

Review the menu whenever headcount changes materially, and again when a line starts producing exceptions. Labor cost is the other moving input, and it moves by market. The BLS occupational employment and wage tables publish estimates by occupation, industry, state and metro area, which is how you check whether a staffing assumption still holds for a line you sell.

The Labor Department's page on misclassification myths explains that the economic reality of the relationship decides worker status, which matters on specialist lines staffed with occasional contractors.

The decisions above the menu, delivery model, capacity and opening sequence, are set out in the startup and market guide.

Common questions

Should I offer everything a client asks for?

Not at the point of asking. Take it as a request, run it through the four axes and the retirement test, and answer within a day. A service accepted in a meeting is a service designed by accident.

Can one assistant cover inbox work and bookkeeping support?

They can, but not in the same hours, and the access requirements differ enough that you should decide the separation deliberately rather than let it emerge.

How many service lines is too many?

When your training material no longer covers everything you sell. That limit arrives earlier than owners expect.

Do specialist lines justify a higher price?

They justify a different price, derived from what they cost you to staff and cover. Whether the market pays it is a question your own win rate answers within a quarter.