Why Your Last Assistant Did Not Work Out

Key takeaways

• Failure one: nothing was written down, so every week was week one and nothing survived a change of person.
• Failure two: nobody was managing them. No coaching, no feedback loop, no cover, and a founder who did not want to be a manager quietly becoming one.
• Only after those two does it become about the person. And only after all three does it become about you.

We hear the same story on the first call more often than any other. "I tried this. I had an assistant for eight months. It was fine, then it was not, and now I do everything myself again."

Then, almost always, a version of: "I think I am just hard to work for."

Sometimes that is true. Usually it is not, and the real reasons are more fixable and less personal. Across the hundreds of these conversations we have had, two failures account for almost everything, and only one of them is about who you hired.

The Story We Hear on Every First Call

It starts well. The assistant is bright and keen. You hand over the inbox and some scheduling. For a few weeks things feel lighter.

Then the small misses begin. A meeting booked over a flight. A follow-up that did not go out. A reply to a client in a tone you would never use. Each one is small. Each one costs you ten minutes to fix and a little trust. You start checking their work, which takes nearly as long as doing it. You stop handing over new things because explaining them feels like a risk. Three months in, they have the inbox and not much else, and you are doing the rest plus the checking.

Then they leave, or you let them go, and because nothing was written down, the next person starts from exactly where the first one did.

If any of that is familiar, here is what actually happened.

If you have been through two or three assistants and none of them worked, the common factor is worth sitting with. It is usually not the one you think.

Failure One: Nothing Was Written Down

The first failure is almost invisible while it is happening, because it looks like normal work.

You explained how you like things done. Verbally, in passing, on a call, in a Slack message that scrolled away. The assistant did their best to remember. They remembered most of it. The bits they did not remember became the small misses, and the small misses became the trust problem.

Nothing was ever captured. Not the ten people whose emails get flagged immediately. Not the day-7, day-14 invoice process. Not the fact that you never take a 9am on a Monday. So every week was, in a sense, week one. The assistant was relearning your preferences from your reactions, and you were re-explaining things you were sure you had already said.

Then they left, and all of it left with them. Eight months of accumulated context, gone, because it lived in a person rather than a document.

This is why we start every client with the Executive Support Checklist and keep everything in the Executive Atlas. Not because documentation is virtuous. Because it is the only thing that makes the second month better than the first, and the only thing that survives a change of person.

Failure Two: Nobody Was Managing Them

The second failure is the one founders feel guilty about, and should not.

An assistant, like anyone, needs a manager. Someone who gives feedback that is specific and early. Someone who trains them when the tools change. Someone who notices they are struggling before the misses start. Someone who arranges cover when they are ill, and who handles the awkward conversation when something is not working.

In a founder-led business that manager is, by default, you. And you did not hire an assistant in order to take on a direct report. So the management did not happen, or happened in bursts when something went wrong, which is the worst kind. The assistant got no coaching, improved slowly if at all, and had nobody to ask when they were unsure except the person they were most afraid of disappointing.

Here is the uncomfortable version: a good assistant placed with a founder who cannot manage them produces a frustrated good assistant and a founder who concludes assistants do not work. The failure looks like a hiring failure and it is almost never a hiring failure.

This is the part of the job Atlas keeps. Your operator has an account manager who owns the relationship, monthly coaching, a pod of other operators to learn from, and cover when they are away. You get the Friday wrap and a daily note. You do not get a direct report.

The Third Thing, Which Is About You

Only once those two are in place does the question of fit arise, and only after that does it become about you. But it does become about you, and this is the part most articles skip because it does not sell anything.

Delegation is a skill with a predictable progression. You start by handing over tasks. Then projects. Then recurring workflows. Then decisions. Then outcomes. Each step requires more trust than the last, and more willingness to let something be done differently from how you would do it.

Most founders stall between workflows and decisions. They will hand over the inbox but not the judgment about which emails matter. They will hand over the calendar but approve every meeting. At that point the assistant is executing and the founder is still deciding, which means the founder is still the bottleneck, just with an extra step.

If you have been through two or three assistants and none of them worked, the common factor is worth sitting with. It is usually not the one you think. It is usually the stall point, and it is fixable, because it is a skill and not a trait.

What Good Looks Like Instead

When this works, four things are true, and they are true in this order.

1. The work is written down. Every recurring task has a trigger, a standard, what done looks like and when to escalate. It lives in a document you own.
2. Someone else is managing the person. Coaching, feedback, training, cover. Not you.
3. The person was selected and trained for this specific job. Not an administrator who happens to be available. Someone screened for judgment under ambiguity and trained on your systems before day one.
4. You are handing over areas, not tasks. The inbox, not individual emails. Travel, not individual bookings. And you are moving up the progression on purpose.

Notice that three of the four have nothing to do with who you hire. That is the point.

How to Tell Which One Happened to You

Go through the last time it did not work and answer these.

• When they left, what documentation existed of how you liked things done? (Nothing: failure one.)
• Who gave them feedback, and how often? (You, when something went wrong: failure two.)
• Who covered when they were ill or on holiday? (You: failure two.)
• Were they trained on your tools and systems before they started, or did they learn on you? (On you: failure two, and partly a selection problem.)
• Six months in, what were they deciding without you? (Nothing: the stall point, and that one is yours.)

Most people get three or four of these. Almost nobody gets zero, which is why "I am just hard to work for" is almost always the wrong diagnosis. You were running a structure with two load-bearing walls missing and blaming the roof.

If You Are About to Try Again

Do three things differently, in this order.

Build the checklist before you hire anyone. Forty minutes from your calendar and sent folder. Knowing what you are handing over changes who you look for.

Decide who manages them, and if the answer is you, be honest about whether you will. If you will not, choose a model where someone else does.

Then, and only then, choose the person. For what to look for and how we do it, this is how Atlas selects and trains operators. For what the first hundred days look like when all three are in place, start here.

Frequently Asked Questions

How Do I Know It Was Not Just a Bad Hire?

Ask what was written down when they left. If the answer is nothing, and nobody was coaching them, you cannot know whether the person was the problem, because the two things that let a good person succeed were missing. Fix those first and the next hire will tell you.

I Am Not a Manager and I Do Not Want to Be One. Is This for Me?

That is exactly who the managed model exists for. Your operator is coached, covered and managed by Atlas. Your part is ten minutes a day and a Friday wrap, which is not management, it is being a good client.

What If I Was the Problem?

Then this is good news, because you are the one variable you can change. Delegation is a learnable progression, and the Energy Audit and readiness coaching exist because most founders stall at the same point in it.

Does the Rematch Really Carry the Context Across?

Yes. Your Executive Atlas, the checklist and every written process belong to you, not to the operator. The next person starts from the manual, not from a blank page.

---

Atlas places and trains executive operators, and coaches the founders who work with them. See what you are actually buying or book a call.