How I onboard and train a new VA in a week
# How I onboard and train a new VA in a week
Most people who tell me hiring an assistant did not work for them describe the same month. Tasks keep coming back half done, they find themselves redoing work at eleven at night, and they quietly conclude that getting help was a mistake. Nearly every time, the hire was fine and the onboarding was the problem. Someone got handed a login and a vague list of jobs, wished good luck, and left to guess.
Here is the process I run instead. It takes about a week, none of it is complicated, and most of it amounts to going slow on purpose in the right order.
Week one is an investment, and it will feel slow
The first week with a new assistant feels inefficient. You will spend an hour handing over something you could have done yourself in ten minutes, and part of your brain will keep insisting you are wasting time. You are buying months. Every hour of clarity in week one is an hour you never spend re-explaining the same task in week nine.
I learned the other path first. When I rushed someone through their first days to save myself a little time, I paid for it across the next quarter: the same task explained again and again, the same small mistakes fixed repeatedly, and trust slowly eroding on both sides. Clarity at the start is the cheapest insurance you will ever buy. Rushing at the start is a bill that arrives later, with interest.
Day one: access and expectations, nothing else
The goal on day one is to get the person set up, comfortable, and clear on how you will work together. Real output can wait.
Access comes first, and I give scoped logins, never my master passwords. Wherever a tool supports inviting a team member with limited permissions, I use that. If something goes wrong, or the day comes when we part ways, I want to switch off one login cleanly instead of changing every password I own in a panic.
Then the tools, and only the two or three they will actually touch in week one. Dumping the whole stack on someone on day one buries them. Then the rhythm: where we talk, how fast they can expect answers from me, and when I am usually offline, so they are never stuck and afraid to ask a small question. And I ask how they like to work.
A calm first day is the entire objective.
Record then document
This is the engine the whole week hangs on. The first time I hand off a task, I skip the wall of text and the live walkthrough where somebody scrambles to take notes. I do the task once myself, on a screen recording, talking out loud through every step. Why I click this. What I check here. What I do when this one field looks off.
The recording costs me one extra pass through a task I was doing anyway, and it becomes the raw material for a simple written procedure the assistant can follow without me. The recording shows how the task feels and flows. The written guide gives them something solid to check against, step by step, without scrubbing back through video every time. Once both exist, the task has an answer that lives outside my head. That is the whole shift. The knowledge stops being mine alone and becomes something we both own.
Where AI earns its place
Writing that first guide from scratch is tedious. The raw transcript of a screen recording is full of false starts and me saying hang on, let me do that again. So I hand the transcript to an AI model and ask for a clean first draft of the procedure: numbered steps, plain language, right order. It is genuinely good at exactly that, and it turns half an hour of typing into a starting point that appears in seconds.
The draft is never the finished thing, though, and this part matters. The assistant and I read it together, fix what the model quietly got wrong, and add the small warnings that only come from having actually done the task. The person makes the guide their own. The AI does the boring first pass, and a human owns the final version every time. I am using it to skip the typing so our attention goes to the part that needs a person, which is getting it right.
One task, then the next, never the whole job
The classic mistake, and I made it too: you are so relieved to finally have help that you hand over everything at once. Start with one task instead. One small, well defined slice of work with a clear finish line, so they always know what done looks like.
Watch the first few runs closely. You are there for feedback, and the feedback should be kind and specific at the same time. Saying that was great teaches nothing, and saying that was wrong teaches resentment. Useful feedback points at the exact spot. This part was right, keep doing it this way. That field needs the date in a different format, and here is the reason why.
Once they can run the first task without you looking over their shoulder, widen. Add the next task, then the next, and give each one the same treatment: record it, write the guide, watch the first few runs, let go. Before long you are handing over whole areas instead of single tasks, because the person has learned how you think, which goes further than knowing which buttons to click.
The feedback loop that makes it stick
In the first week I do a short daily check-in, five or ten minutes over chat. What did you work on. What got in your way. What was unclear. That is the whole agenda. I am closing the gap between a small mistake and someone catching it, so nothing hardens into a bad habit.
As things click, I back off on purpose. Daily becomes every couple of days, then twice a week, then whenever we actually need it. The point of the arrangement is to need each other less over time.
And the honest part, which took me far too long to accept: almost every mistake in week one traces back to an unclear instruction, and unclear instructions are mine. When something comes back wrong and I feel the frustration rising, the real fix is almost never a stern conversation. It is me rewriting a guide so the next person cannot misread it.
The library that compounds
Every task you record and document becomes a permanent asset, and a small library of how your business works starts to grow, written in plain language, somewhere you can both reach it. The first guide is real work to produce. The tenth is far easier, because a template and a rhythm exist by then.
Then a line gets crossed that changes the whole thing. The assistant stops just following the guides and starts improving them: fixing a step that reads unclear, documenting a faster method they found. The library gets better without you touching it. That is the moment a hire quietly turns into a teammate.
The knowledge is written down and shared, and it gets a little sharper every week. If the person ever moves on, the library stays. The next hire starts from everything the last one learned instead of starting from zero.
What the end of week one looks like
One week is short, and I am careful about what I promise myself from it. I am never expecting someone who can run the whole business after seven days. That takes months. What I expect is a person who can run two or three specific tasks fully on their own, a small set of written guides covering exactly those tasks, and a working relationship where they tell me when something is unclear instead of guessing and hoping.
That is a strong first week. A few tasks off your plate for good, documented so they stay off, and a person who keeps getting better from here. The final day never looks dramatic. It mostly looks like more breathing room than you had seven days earlier, and a foundation the next month can build on.
If you have not hired yet
The onboarding decides whether the hire works, and now you have the playbook: record then document, one small task at a time, daily check-ins that taper, a library that compounds. Finding the person is its own topic, and I hire through OnlineJobs.ph myself. My full hiring process, and the rest of how I run a small business with remote help, is on the channel and at [xavierfok.com](/).
Get new guides and videos first — join the Telegram channel.