How to Vet a Remote Executive Assistant: The Work-Sample Test We Use
To vet a remote executive assistant, score a short written application and an English test, hold an unscripted video call, then give the best candidates a paid work sample that looks like the real job. Fail anything that doesn't work, score the rest on a fixed 25-point rubric, and interview only those who pass, with the same questions for everyone. Then check references, verify identity on camera and start with a paid trial week.
Vetting a remote executive assistant at a glance
| Step | What you're testing | Time it takes you |
|---|---|---|
| 1. Written application | Effort, writing, ambition | 5 minutes per applicant |
| 2. English check (CEFR) | Can they run your inbox and calls in English? | 2 minutes to read a score |
| 3. Live video call | Real spoken English, presence, situational judgment | 15 to 20 minutes |
| 4. Paid work sample | Can they do the actual job? | 30 minutes to brief, 30 to score |
| 5. Structured interview | How they think about the work they just did | 30 minutes |
| 6. References | What they built for someone else | 15 minutes per call |
| 7. Identity check | Is this the person who'll do the work? | 5 minutes |
| 8. Paid trial week | Do they finish real tasks, on time, without you? | One week |
I'm Ivan, co-founder of Pareto Talent. I started as an executive assistant myself, and since 2024 we've tested 36,186 people who applied to work with US founders. 518 of them passed our final work test. This guide is that vetting process, cut down so a founder can run it alone for one hire. It works for any remote hire, but the work samples and the rubric are written for executive assistants.
01 · The Process
How do you vet a remote executive assistant?
You vet a remote executive assistant by watching them do the job before you hire them. Résumés and charming interviews tell you very little. A realistic, paid work sample, scored on a rubric you wrote in advance, tells you most of what you need. Everything else in the process exists to decide who gets the work sample, and to confirm what it showed.
Here's the order, and why it's in that order:
- Cheap filters first. The written application and English check cost you minutes. They remove most people.
- Expensive filters last. The work sample and interview cost you real time, so only a shortlist gets them.
- The interview comes after the work, not before. You'll ask about what they built, which is much harder to fake than talking about yourself.
The research agrees. In the most recent large review of hiring methods, structured interviews, job knowledge tests and work samples predicted job performance best. Years of experience barely predicted it (Sackett et al., 2022).
What actually predicts job performance
Correlation between each hiring method and later job performance (0 = none, 1 = perfect). Updated 2022 meta-analysis.
- Structured interview0.42
- Job knowledge test0.40
- Work sample test0.33
- Cognitive ability test0.31
- Unstructured interview0.19
- Years of experience0.07
Source: Sackett, Zhang, Berry & Lievens (2022), Journal of Applied Psychology 107(11), operational validity estimates.
02 · The Application
What should a remote executive assistant application ask?
Ask every applicant for the same things so you can compare them fairly: a CV in English, written answers to two or three questions, an English test result and a two-minute recorded video introduction. Score each part from 1 to 5 and decide your cut-off before you start reading.
The three questions we've found most useful:
- "What's the biggest project you've ever owned from start to finish? What was the result?" You want something specific with a number in it.
- "Tell me about a mistake you made at work and what you changed afterwards." You want ownership, not blame.
- "Where do you want to be in three years?" In our data, people who pass describe bigger, more entrepreneurial goals. Ambitious assistants are the ones who build systems for you.
Expect most people to drop out here, and treat that as a feature. Of everyone who signed up with us, only 16% completed the full application. Someone who won't finish a 20-minute form won't finish your projects.
Two things to watch. First, answers written entirely by AI: polished, generic and free of any real detail. Use of AI isn't the problem; using it without thinking is. Ask a follow-up about a specific detail and you'll see quickly. Second, the career behind the application. In our bootcamps, former freelancers (28%) and marketers (27%) passed the final test at almost three times the rate of former call-center agents (10%). People who have shipped their own work recognize the job. People who've followed scripts often need more than a script.
Builders pass, scripts don't
Share of bootcamp participants who passed the final project, by previous career. Average across everyone: 16%.
- Former freelancers28%
- Marketers27%
- All participants16%
- Former call-center agents10%
Source: Pareto Talent applicant data, Bootcamps 2 to 7 (2024 to 2026). See method.
03 · English
How good does a remote executive assistant's English need to be?
For an assistant who writes to your clients and joins your calls, aim for C1 on the CEFR scale (advanced). B2 can work for back-office tasks. In our data, the pass rate on our final work test climbs from 6% at B1 to 11% at B2 and 16% at C1, then barely moves at C2 (17%). 72% of the people who pass test at C1 or above.
English matters up to C1, then it flattens
Bootcamp participants who passed the final project, by English test level (CEFR).
- B1 (intermediate)6%
- B2 (upper intermediate)11%
- C1 (advanced)16%
- C2 (proficient)17%
Source: Pareto Talent applicant data, Bootcamps 2 to 7 (2024 to 2026). See method.
Ask for a recent result from a standard online test that reports a CEFR level, or run one yourself. Then check it live on the video call. A written score and a real conversation can differ, and the conversation is what your clients will hear.
04 · The Video Call
What should you check on the first video call?
Hold a 15 to 20 minute unscripted call in English, camera on. Don't pitch the job for 15 minutes. Let them talk. You're checking three things: can they hold a real conversation at speed, do they think before they answer, and do they ask you good questions.
Use two situational questions, the same ones for everyone. These come from our own interview guide:
- "You're managing my inbox and an urgent email arrives while you're in the middle of something else. What do you do?"
- "I ask you to learn a tool you've never used, by tomorrow. How do you approach it?"
Good answers are specific and have an order: what they'd do first, what they'd tell you, what they'd do next. Weak answers are general ("I'm very organized") or ask you to decide everything. Score each answer 1 to 5 and move the top three to five candidates forward.
05 · The Work Sample
What is a good executive assistant skills test?
A good executive assistant skills test is a small, real piece of the job: paid, with a deadline, built from a sanitized copy of your own work. Give 3 to 5 hours of work with 48 to 72 hours to finish, and ask for every deliverable as a shared link that you can open.
Don't use quizzes about software. Use your actual work. Here are three briefs you can copy. Pick the one closest to the job, or combine two into a bigger test.
Brief A: Inbox triage. Export 30 to 40 emails from a real week (remove anything sensitive, or rewrite them) into a shared folder or a test inbox. Ask the candidate to: sort every email into reply now, delegate, schedule or archive; draft replies to the five most important in your voice; write a five-line summary of what needs you today; and set up labels and two filter rules that would handle this kind of week automatically. Deliverables: the sorted list, the drafts, the summary, and a screenshot or screen recording of the working filters.
Brief B: Plan my week and a trip. Give them a copy of next week's calendar, a list of six meetings that need to fit in, your rules (no calls before 10, two focus blocks a day, a gym slot) and a two-night trip to another city with a budget. Ask for: a proposed calendar with every conflict resolved; two flight options and one hotel inside the budget, with reasons; a one-page itinerary; and the emails that would move or confirm each meeting. Deliverables: a shared calendar or document, the itinerary and the email drafts.
Brief C: Build a simple follow-up automation. Ask them to build, in tools you already use or free versions of them: a short form that collects a lead's name, email and what they need; an automation that adds each submission to a sheet or CRM, sends a confirmation email and creates a follow-up task two days later; and a test with three fake submissions. Then a two-minute screen recording explaining how it works and what would break it. Deliverables: the live form link, the sheet or CRM view, and the video.
Pay for the work at the candidate's expected hourly rate. It's fair, it gets you their real effort, and it shows strong candidates you take their time seriously. Tell them up front that you'll score it on a rubric, and that they may use AI tools as long as they can explain the result.
The rule that does the most work: every deliverable must work
Before you score anything, open every link. If a link is broken, a form doesn't submit or an automation doesn't fire, the work sample fails. No partial credit. That sounds harsh. It's the point. In real work, an email that almost went out or a booking that almost got made is a failure your client sees. We run the same gate on our own final project, and it filters for the one habit you can't train in: checking your own work.
The second filter is simply finishing. In our latest three cohorts, only 36% of people who started our training submitted a final project, and 38% of submitted projects passed. Some of your best-looking candidates won't send anything. That tells you something too.
Source: Pareto Talent Bootcamps 5 to 7 (2025 to 2026) for the first two figures; Bootcamps 2 to 7 for the total. See method.
06 · The Rubric
How do you score an executive assistant work sample?
Score it on five categories, 1 to 5 points each, 25 in total, and write the rubric before you read a single submission. This is the same structure as the Pareto AI Operator Scorecard we use on our own five-day build, with the categories rewritten for an executive assistant.
| Category | What a 5 looks like | What a 1 looks like |
|---|---|---|
| Judgment & priorities | The right things come first. They flag what needs you and handle the rest. Assumptions are written down. | Everything treated as equally urgent, or everything sent back to you. |
| Writing & tone | Drafts you could send as they are. Short, clear, in your voice. | Long, stiff or generic. You'd rewrite every line. |
| Accuracy & detail | Names, dates, time zones and amounts all correct. Nothing missed. | Wrong dates, missed conflicts, broken details. |
| Tools & systems | Uses the tools well and sets up something that keeps working without them (filters, a template, an automation). | Does everything by hand, once. |
| Communication & presentation | A clear summary or video: what they did, why, what's left and what they'd ask you. | Delivers files with no explanation, or explains at length without saying anything. |
Set the bar before you start. The thresholds we use on our own project: under 15 out of 25 is a no. 15 to 19 is "promising, not yet". 20 or more moves on to the interview. One person scoring every submission, against the same rubric, in one sitting, keeps it fair.
Want candidates who've already passed this test?
Every Right Hand we match has passed our application, 10 days of live training and a scored five-day build. Tell us what you need and see three hand-picked executive assistants within 24 hours.
07 · The Interview
What executive assistant interview questions should you ask?
Ask questions about the work sample they just did, the same ten for everyone, in the same order, and score each answer from 1 to 5. Structured interviews like this were the single best predictor in the 2022 review, about twice as predictive as the casual, unstructured kind (0.42 vs 0.19).
Ten questions, each mapped to a rubric category:
- Judgment: "Which email in the batch did you put first, and why that one?"
- Judgment: "What did you decide not to bring to me? Where's the line for you?"
- Writing: "Read me one of your drafts. What would you change if I said it was too formal?"
- Writing: "How did you work out my tone without ever having talked to me?"
- Accuracy: "How did you check your work before you sent it?"
- Accuracy: "Walk me through the time zones in the trip. What could go wrong?"
- Tools: "What did you set up that would keep working next week without you?"
- Tools: "Did you use AI anywhere? Show me the prompt and what you changed in the output."
- Communication: "What did you not have time to finish, and how did you decide what to cut?"
- Communication: "If this were your first week with me, what would you ask me in our first daily check-in?"
Add two alignment questions at the end, unscored: their pay expectation, and when they could start. Then give them five minutes to ask you anything. The questions they ask tell you how they think about the job.
08 · References
How do you check references for a remote executive assistant?
Call two references, preferably people they worked for directly, and ask about output, not personality. "Were they nice to work with?" gets a yes every time. These questions get real answers:
- "What did they own for you, start to finish?"
- "What's one thing they built or set up that you still use?"
- "How did they handle a mistake?"
- "Would you hire them again for the same role? Why or why not?"
Listen for hesitation on the last question. A reference who pauses has told you something.
09 · Identity
How do you verify a remote candidate is who they say they are?
Meet on camera at every live step, ask to see a government ID on the video call before you make an offer, and check that the face, the name and the person who did the work sample all match. Do it before you share a single login.
This isn't paranoia. In a 2025 Gartner survey of 3,000 job candidates, 6% admitted to interview fraud: posing as someone else, or having someone else pose as them. Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake (Gartner, 2025). An assistant gets your inbox, your calendar and often your cards, so the identity check matters more here than for most hires.
Two practical habits: keep the camera on in the work-sample walkthrough as well as the interview, and give access in levels during the trial, starting with draft-only email and no payment tools.
10 · The Trial
Should you do a paid trial week?
Yes. A paid trial week of real tasks is the last and most honest test. Give your top candidate the work you'd give them in week one, a 10-minute daily check-in, and watch three things: do they finish on time, do they ask good questions early, and do they tell you about problems before you find them.
Pay the trial at the full rate and agree the end date in writing. If you have two strong finalists, run the trial with both in parallel on similar tasks. One week of real work will settle what three interviews couldn't.
Red Flags
What are the red flags when vetting a remote assistant?
- The work sample doesn't arrive, or arrives late without warning. Finishing is the job.
- A broken link or a form that doesn't submit, and they didn't notice.
- Polished, generic writing that falls apart when you ask one follow-up question.
- Camera always off, or a different voice or face from one call to the next.
- They can't explain their own work in plain English, especially work done with AI.
- Every question comes back to you. An assistant who needs a decision for everything won't save you time.
- References who are friends, or who can't name one thing the person built.
- Pressure to skip steps: "I can start tomorrow, no need for a test."
Checklist
The remote executive assistant vetting checklist
Copy this into your hiring doc.
1. Write the 90-day outcome for the role before the job post.
2. Application: CV in English, three written answers, an English result and a two-minute video. Score 1 to 5 each.
3. English: C1 for client-facing work. Confirm it live.
4. Video call: 15 to 20 minutes, camera on, two situational questions, scored.
5. Work sample: paid, 3 to 5 hours, 48 to 72 hours to deliver, every deliverable a working link.
6. Gate: one broken deliverable fails it. Then score 25 points; 20 or more goes forward.
7. Interview: the same ten questions about the work sample, scored 1 to 5.
8. References: two, about output, not personality.
9. Identity: government ID on camera before any offer or login.
10. Trial: one paid week of real tasks, daily 10-minute check-in, access in levels.
For the wider picture, including where to recruit, what each hiring model costs and what our data says about which countries and careers produce the strongest candidates, read the full guide to hiring remote talent in Latin America and the Latin America Remote Talent Report.
FAQ
Frequently asked questions
How do you vet a remote candidate?
How do I hire a remote executive assistant?
What is the best executive assistant skills test?
What questions should I ask in an executive assistant interview?
What English level does a remote executive assistant need?
How long should a remote interview process take?
How do I avoid hiring a fake remote candidate?
Data Notes
About the data in this guide
Our data covers everyone who applied to Pareto Talent's training program for remote executive assistants and AI operators, Bootcamps 2 to 7 (2024 to 2026): 36,186 people. 5,840 completed the full application, 3,250 joined a live bootcamp and 518 passed the final project, a five-day build scored on a 25-point rubric. Passing the final project is not the same as being placed with a client. Attendance and submissions were tracked for Bootcamps 5 to 7 only, so the 36% and 38% figures use those cohorts. Careers were classified from applicants' own written answers. Differences are associations, not causes. Full method: Latin America Remote Talent Report.
External data is linked where it's used and listed below. The work-sample briefs, the executive assistant rubric wording and the interview questions are our own recommendations, adapted from the process we run.
Sources
- Pareto Talent, Latin America Remote Talent Report 2026, applicant data 2024 to 2026.
- Pareto Talent, The Pareto AI Operator Scorecard.
- Sackett, Zhang, Berry and Lievens (2022), Revisiting meta-analytic estimates of validity in personnel selection, Journal of Applied Psychology 107(11).
- Gartner (31 July 2025), Gartner Survey Shows Just 26% of Job Applicants Trust AI Will Fairly Evaluate Them, press release (2Q25 survey of 3,000 job candidates).
Meet three pre-vetted executive assistants in 24 hours
Every Right Hand we match has already passed the application, the English check, 10 days of live training and a scored five-day build. Tell us what you need and we'll show you three hand-picked profiles. If we don't find a match you're truly excited about, you owe nothing.