Hiring again after a bad VA experience: what to change the second time

About a fifth of the agencies that come to us have already tried this once and it went badly.

They're not vaguely skeptical. They're specific. They lost months, they lost money, some of them lost a client. And they arrive with a reasonable instinct: this time, be more careful.

The problem is that "more careful" usually means scrutinizing the same things harder. And the things most owners scrutinize were never what went wrong.

Our co-founder, Dominic Piccirillo, has written about why these arrangements fail, and the short version is that the failure is structural rather than personal. This piece is about what to do with that. I've made hundreds of placements for independent agencies and built our intake process from scratch, refining it placement by placement, so what follows is what I'd want you to change before you try again.

The three ways the second attempt repeats the first

1. You screen the person harderMore interviews, longer trials, a closer read of the résumé. It feels like diligence. But a résumé told you nothing useful the first time and it won't this time either, because "virtual assistant" is a job title rather than a capability. Meanwhile the variables that actually predict the outcome, how the match was made and what support sits behind it, still aren't being examined.
2. You de-risk by spending lessThe most common move and the most costly one. Having lost money on the first attempt, the instinct is to limit the downside on the second. So you go cheaper, which means further down-market, which means a marketplace optimizing for who's available at the lowest rate. That's the model that failed you. Paying less for it makes failure more likely, not less expensive.
3. You shrink the scope to almost nothingYou hand off one small, safe task, keep everything else, and wait for it to go wrong. It usually doesn't, because the task was trivial. It also doesn't help, because you're still doing the work that was burying you. Six months later the arrangement quietly ends, having proved nothing except that you can delegate something that didn't matter.

Each of these feels like caution. All three preserve the conditions that caused the first failure.

Evaluate the model, not the candidate

Here's the reframe that changes the outcome.

The candidate matters, obviously. But you have very limited ability to assess a stranger across two video calls, and you know that already, because you tried it. What you can assess in detail, before committing anything, is the provider's model. That's the variable with the most predictive power and the one almost nobody examines.

The red flags, specifically

These are what I'd tell any agency owner to benchmark against, whoever they're evaluating.

• No insurance-specific trainingNot "we train our people" in general terms. Ask whether they've worked inside your management system before they touch your account.
• No personality or culture fit assessment, just a résumé matchIf nobody asked how your agency actually runs before proposing a name, nothing has changed since last time.
• No dedicated support after placementA login and good luck means onboarding and management quietly transfer to you, the person who had no time to begin with.
• No visibility into what the assistant is working on day to dayYou should not have to ask what happened this week.
• Overnight work with no real-time overlapA twelve-hour offset means every question becomes a next-day item, and permanent night shifts show up eventually as turnover.
• Part-time or project-based rather than a committed team memberShared pools work for one-off tasks. They fall apart when the work needs context.

If two or more of those describe what you're being offered, you're looking at the same arrangement in different packaging.

The questions worth asking

Take these to every provider, including us, and compare the answers rather than the pricing.

1. On what basis did you select this candidate for my agency specifically?
2. What is this person trained on before day one, and what training falls to me?
3. Who handles onboarding, and what does the first thirty days look like?
4. What happens if the fit is wrong? How fast is a replacement, and what does it cost me?
5. Which hours will they work, and on whose clock?
6. Who do I contact when something breaks, and what's the response time?
7. What percentage of your placements are still in place a year later?

That last one is the question I'd want asked of me. A provider who churns through replacements is telling you their matching doesn't work, and you'll experience that as your problem rather than theirs. Ours sits at 91% retention across all placements, and the reason isn't the talent pool. It's the questions asked before anyone is ever matched.

Watch how the answers come, too. Hedging on the matching question or the replacement question is the signal worth paying attention to.

Do your own homework before the first call

Half of a good match is information the provider can't get without you. Before you talk to anyone, have answers to these.

• What tasks will the assistant handle on a typical day?
• What AMS or CRM do you use, and how well documented are your workflows?
• Who manages the assistant day to day, and what's their actual bandwidth?
• What's your management style: hands-on or autonomous?
• Will they handle inbound calls, and if so what type and volume?
• What's your agency's culture and pace: fast and scrappy, or steady and process-driven?
• Do you need bilingual support?
• What does success look like at 30, 60, and 90 days?
• What time zone coverage do you need?
• What failed in your past hiring, and what would you do differently?

That last question is in our intake form for a reason. When an owner tells me exactly what broke last time, it changes which candidates I even consider. An owner who says "work bounced back to me the moment anything went off-script" is telling me they need someone with enough insurance grounding to make a judgment call, not just someone who follows instructions. That's a different person entirely.

So when I say tell your next provider what went wrong, I don't mean it as good manners. It's the single most useful input you can give, and how a provider uses it tells you whether their matching is real.

Set it up before day one

Most placements fail not because of skills but because the agency wasn't set up to integrate the person. This is the part owners underestimate, and it's mostly done before the assistant ever logs in.

1. Have the tech stack readyZoom, VOIP, and AMS access provisioned before they start. Not on day three. An assistant spending their first week waiting on credentials is a week you paid for and lost, and it sets the tone that they're peripheral.
2. Put them in the daily huddle and the team channels from the startNot a weekly check-in with you. The actual rhythm your team runs on.
3. Write down your 30/60/90 expectations in advanceObservable things, not feelings. At thirty days, is the workflow moving without you in it? At sixty, is it going off-script less often? At ninety, have your licensed people reclaimed hours, and can you name where those hours went? Bad arrangements rarely end in a decision. They fizzle. Written checkpoints force the conversation while there's still time to correct.
4. Include them in new business and retention bonuses, the same as your US staffThe one owners are most surprised by, and the highest-leverage thing on the list. An assistant who shares in the outcome behaves like someone who shares in the outcome. They flag the cross-sell, they chase the renewal, they tell you what they're seeing. Treat them as a vendor and you'll get vendor behavior, which is exactly what you're trying to escape.
5. Give real work, not safe workThe instinct after a bad experience is to protect yourself with something low-stakes. Do the opposite, within reason. Hand over a real recurring workflow with actual volume, ideally the one eating the most of your licensed staff's time. You'll learn more in two weeks than in six months of trivial tasks, and if it isn't working you'll know early enough to fix it.

When to walk away, and when it's fixable

Not everything that feels wrong in week three is a failed placement.

How a provider responds to a problem in week three tells you most of what you need to know about the next three years.

Find out what that replacement process is before you sign, not while you're in it.

How we do it, in the open

I've spent this whole article telling you to evaluate the model. It's only fair that I show you ours.

Every match I make weighs four things: workstyle compatibility, technical strengths, shared values, and communication style. Availability is not one of them. That's the whole reason the 91% number exists.

On training, our assistants complete a four-to-six-week insurance program before placement, including hands-on work inside HawkSoft, insurance terminology, and dedicated lessons in personal and commercial lines workflows. Peggy Corbett, a thirty-year agency owner and HawkSoft Master Trainer, runs a session with every cohort. Only half our trainees graduate the program, and 0.6% of applicants are hired.

The process itself: a call with me, a matchmaking intake form, interviews with your top two or three matched candidates, and a start date usually two to three weeks out. You meet people who were selected for your agency, not people who happened to be free.

And for agencies without documented processes, which is most of them, Elevated+ builds the workflows alongside the placement, so the assistant arrives trained on how you actually work rather than on a general idea of insurance.

None of that makes us the only reasonable choice. It does make every red flag on that list answerable, which is the bar I'd hold any provider to. Here's how the matching works in detail.

Your first attempt proved less than you think

A bad experience feels like data about whether this works. It's really data about one model: the one that matched on price, skipped the training, and left the managing to you.

The agencies that get it right the second time aren't the ones who got luckier with a candidate. They're the ones who changed what they were evaluating.

This time, do it differently.

Book a call and tell me exactly what went wrong last time. I'll be straight with you about whether we'd do it differently, and how.

Get Matched