Every growing team hits the same wall. The work keeps arriving, the calendar keeps filling, and the same four people carry all of it. Revenue is up, sleep is down, and the question lands on the owner’s desk: do I hire, automate, or outsource this?

Most owners answer with whatever feels fastest. A job post goes up. Or a tool gets bought and never used. The task keeps landing on the same desk.

I founded a marketing agency, grew it for eight years, and sold it. I now teach marketing and communications teams to use Claude and Claude Code without a developer, and I help owners decide how work gets done. The decision comes down to a short set of questions, not a feeling.

Start With the Task, Not the Headcount

Owners usually frame the problem as a people problem. It is a task problem. Break the work into named tasks before you decide anything. “Marketing” is not a task. “Write the weekly newsletter, edit it, schedule it, and pull the open rate” is four tasks, and each one has a different answer.

Write the tasks in a list. One line each. Keep verbs in front. If a task takes less than an hour a week, it stays in the list. Small tasks are where the wall is built.

Three Ways to Get a Task Done

A hire buys judgment, context, and time. You get a person who learns the business, spots problems you did not name, and carries work across many tasks. A hire costs the most and takes the longest to pay off.

An automation buys repetition. A script, a template, or an AI workflow does the same steps the same way every time. It costs the least to run and breaks when the work changes shape.

An outsourced or fractional team buys capacity without a long commitment. You rent skill for a defined scope and stop when the scope ends. It costs the most per hour and the least per decision.

The Test for AI

Run each task through four questions. If you answer yes to all four, AI takes the task. If you answer no to any one, a person keeps it.

  • Does the task start from material that already exists, such as notes, a transcript, a brief, or last quarter’s numbers?
  • Can you describe the output shape in one sentence, such as a 200 word summary or a table of counts?
  • Is a wrong first draft cheap to fix, meaning a person reviews before anything ships?
  • Does the task repeat on a schedule, so the setup pays for itself?

AI drafts. A person decides. That line stays fixed in every workflow I set up. A first draft takes minutes. The judgment call on what goes out the door still belongs to a person who knows the business.

The tasks that fail the test share a shape. They need a relationship, a negotiation, or a call made with half the facts. They touch money, legal exposure, or a client’s trust. They change every time. Those tasks go to a person, and often to a senior one.

Choose Between a Hire and an Outsourced Team

Once the tasks are sorted, count the ones that need a person. Then ask two questions about that pile.

  • Does the pile grow every month, or is it steady?
  • Does the work need context that only lives inside your company?

A growing pile that needs inside context points to a hire. A steady pile with a clear scope points to a fractional or outsourced team. A pile that spikes around launches or reporting cycles points to a contractor you bring back when the spike comes.

Hire when the person will own a result over a long stretch and make decisions you cannot script. Outsource when the deliverable is defined and the clock is known. Automate when the steps are the same every time.

The Cost You Are Actually Comparing

Owners compare a salary to an invoice and stop there. Compare the full load instead.

A hire carries recruiting time, onboarding, tools, management attention, and the months before the person is fully useful. An outsourced team carries a higher rate and no long tail. An automation carries setup time and a maintenance habit.

The honest comparison is cost per finished task in the first ninety days, then cost per finished task after the work settles. Early costs favor the outsourced team and the automation. Late costs often favor the hire, once the pile is big enough to keep one person busy.

Put the Decision in Writing

Write one page. List the tasks. Mark each one hire, automate, or outsource. Note who reviews the AI output and what triggers a change.

A short written decision does three things. It stops the same task from landing on your desk every week. It gives the person you hire a clear scope. It tells you when to revisit the call, such as when the pile doubles or a task stops repeating.

Review the page once a quarter. Work changes shape. So should the answer.

The wall is not a sign that your team is failing. It is a sign that the work outgrew the old split between people and tools. Sort the tasks, run the test, and put the decision in writing.

If you want a second set of eyes on where to hire, automate, or outsource, tell me through the contact page with the one task that keeps landing back on your desk. I read every reply. You can also find the frameworks I use at https://www.unicornflames.com/.