I've hired three part-time assistants over the past four years, and I got the first one badly wrong. I posted a vague job ad, hired whoever interviewed best, and spent two months explaining tasks I could have finished myself in ten minutes.
The next two hires went completely differently because I flipped my approach: I wrote the work down first, hired off a paid test instead of a smooth interview, and locked down access before day one. If you want to hire a part-time assistant without burning a month on the wrong person, this is the method I now follow, step by step.
Part-time assistants handle inbox triage, scheduling, research, travel booking, and post-call follow-ups, typically working 5 to 20 hours per week.
Compared to a full-time hire, the scope is narrower, and the cost is lower. Unlike a task-based assistant service, a single person learns your tools and preferences over time rather than a rotating pool handling isolated requests.
The bad hires I've made started with skipping the prep. Before you post anything, get these four things in order:
{{templates}}
The same process works whether you're figuring out how to hire an executive assistant or a general admin. Only the task list and budget change.
Track one full week of your work before you write anything. I keep a running note and log every task I handle more than twice; last time, this produced 26 items, 12 of which never needed my judgment.
Those 12 became the job description. Vague posts attract vague applicants, and a specific task list filters candidates before you ever read a proposal.
Pro tip: If a task takes longer to explain than to do, keep it. Start with the tasks that follow a repeatable pattern.
Before you write a single word of the job post, nail down your weekly hours, your hourly budget, and whether remote works for you. I start every assistant at 10 hours per week because it's enough to cover a meaningful chunk of the workload without locking yourself into a bad fit. Ten hours gives you enough signal to know whether the relationship is working before you expand it.
Match the budget to the sourcing route you'll choose in the next step, since a freelancer and a managed service fall into different price bands.
While you're at it, pin down worker classification before you start. It's worth a quick check with an HR partner, because getting it wrong ends up costing more than the hourly rate ever would.
Pro tip: Write your budget as a monthly cap. Hours creep, and a cap forces the priority conversation early.
The choice comes down to how much vetting you want to handle yourself:
Pro tip: If you need someone working within two weeks, skip the marketplace and pay the managed premium. Proposal review tends to take longer than you expect.
Paste your Step 1 task list into the post, then add the practical details: hours, tools, time zone, and rate range. I also bury one specific instruction in the middle, like asking applicants to open with the word "cadence", which removes everyone who never read the post.
Keep it short, because applicants skim, and the task list does more to predict fit than anything else in the post. Skip the personality wishlist entirely.
Pro tip: Ask one scenario question, such as how they'd handle a double-booked calendar with both parties traveling. The answer tells you more than any resume line.
Cut the pile in two passes. First, remove everyone who ignored your buried instruction, which in my last round dropped 31 applicants to 9. Then rank the rest against the task list alone. The cover letter polish doesn't tell you much.
Interview the top three for 20 minutes each, no more. I ask which tools they've run for past clients, one scenario from my actual workload, and how fast they turn things around. Twenty minutes is plenty to see how someone thinks, and dragging the call out longer mostly just rewards whoever interviews the smoothest.
Pro tip: Watch how they handle the scheduling of the interview itself. Someone who needs three emails to book a call will probably need three emails for most things.
Don't skip this, and don't make it free. I give them two hours of work straight from my backlog, like cleaning a contact list or drafting replies to five pending emails, and I judge the result against what I would have produced.
The test surfaces things interviews miss. You'll see whether they follow instructions, ask when something's unclear, and meet a deadline without a nudge. A $40 test task has saved me from months of the wrong hire, which makes it the cheapest insurance in hiring.
Pro tip: Give deliberately incomplete instructions once. A good hire asks a clarifying question. Someone who's not a fit guesses and hands you back something confident and useless.
Record yourself doing each recurring task once, with a short screen capture and voiceover, and store the videos in a shared folder. My onboarding doc is seven short videos and one page of preferences, and it got my second hire up to speed in a single week.
Handle access like it can go wrong, because eventually something will. Share credentials through a password manager, grant the minimum each task needs, and schedule a 30-day review to adjust access from there.
Pro tip: Set a daily async check-in for the first two weeks, one message with done, doing, and blocked. Drop it to weekly once trust is earned.
Before my third hire, I went back through my task list with a different question in mind. I sorted it into two buckets: decisions that needed my judgment, and work that was just passing things along. Inbox to calendar, calendar to apps, apps back to inbox.
The second bucket was bigger. Most of what I wanted to hand off was scheduling, triage, and notes, work that doesn't need a person. An AI assistant covers all of it now, for less than I'd pay a human assistant for a single week.
Go human when the work involves:
Try AI first when your list skews toward:
The same logic scales up. Teams that hire assistants for several executives at once can end up paying human rates for admin overflow that software would handle, so run the same split on your list before hiring more people.
If your list is mostly the second group, try an AI assistant for a week before committing to a hire.
{{cta}}
Lindy is an AI assistant you text. It handles your inbox, meetings, and calendar, and connects to hundreds of integrations, including Gmail, Outlook, Slack, HubSpot, and Salesforce.
That admin pile, the scheduling, triage, and notes, is what it was built for.
In practice, it covers the tasks most people hand off first:
You can also set Lindy to text you updates on the things you've asked it to watch, so nothing sits in a queue waiting for you to check in.
Setup is quick, and the Lindy Academy walks you through your first tasks if you need a hand. There's a 7-day free trial with no credit card needed, and paid plans start at $49.99 per month.
Two to four weeks from job post to productive onboarding, in my experience. Managed services cut sourcing time to about a week, but documenting your processes and building trust still take up the remaining weeks.
Documenting your own processes. A lot of people discover their "simple" tasks live entirely in their heads, and writing them down is slower and more humbling than any interview.
It depends on the route. Marketplaces cost the least, managed U.S. services run several times as much, and a part-time executive assistant through a managed service like BELAY sits at the top, at $1,200 to $3,000 per month.
Lindy handles the coordination side well, like email, scheduling, and meeting notes. Anything that requires judgment, relationships, or in-person presence still needs a human. Sort your list into those two categories, and where Lindy fits will be clear.
Honor the 30-day decision date you set before they started. On managed services, ask for a rematch. On marketplaces, end the contract and repost with a sharper job description.

Lindy saves you two hours a day by proactively managing your inbox, meetings, and calendar, so you can focus on what actually matters.
