How we test and score virtual assistant services
Updated
Every score on this site comes from the same process. We sign up as a paying customer, run the same brief through each service, and score what we get. Nothing here is based on a sales demo or a provider’s own claims.
What we test
For each service we:
- Sign up at the standard rate with a real business brief — no press accounts, no freebies.
- Run a four-week brief: inbox triage, calendar management, travel booking and a recurring research task. Where a service specialises (bookkeeping, legal admin), we add a task from that field.
- Track response times, error rates and rework in a shared log, and note every interaction with onboarding and support.
- Re-check prices with the vendor before publishing, and again at least every 120 days. The “Last verified” date on each page is the last time we did this.
How the score is weighted
Scores are out of 10 and weighted like this:
| Criterion | Weight | What it measures |
|---|---|---|
| Quality of work | 40% | Accuracy, judgement and how much rework we did |
| Responsiveness | 20% | Turnaround on tasks and replies across the brief |
| Onboarding and matching | 20% | Time to a productive assistant; how well the match fit the brief |
| Value | 20% | What you get for the effective hourly rate, including minimums |
A score of 8.0 or higher means we would recommend the service without caveats for its “best for” use case. Below 6.0 we do not recommend it; the review will say why.
What scores are not
Scores are our editorial opinion from our own testing. They are not aggregated user ratings, and we never publish an average rating we cannot substantiate. Commissions do not influence them.
Corrections and re-tests
If a service changes materially — new pricing, new ownership, a different assistant model — we re-test and update the score and the date. Spotted something out of date? Tell us.