Every headline says companies are laying people off because of AI.
That’s incomplete, and if you accept it at face value, it leads to the wrong decisions inside your agency.
There are three primary drivers behind these layoffs.
First, AI is improving fast. It is making certain types of work more efficient and, in some cases, reducing the need for specific roles. That is true. Output per employee is going up, and some labor is becoming redundant.
Second is capital. These companies are not just using AI. They are building it. That requires massive investment in infrastructure: data centers, compute, proprietary models. When billions are committed to that buildout, capital has to be reallocated. Labor is one of the fastest ways to do it.
Third, there are always underlying business issues. Bloated teams, inefficient processes, slowing growth. Those problems exist in every cycle. Saying “we are restructuring for AI” sounds a lot cleaner than saying “we overhired and our systems are inefficient.” AI becomes the narrative used to explain decisions that would have been made anyway.
All three are happening at the same time.
Most agency owners see the first driver and stop there. They assume the play is to cut headcount because AI is replacing people.
That is not the situation for most agencies.
You are not building AI infrastructure. You are renting it. The cost structure is fundamentally different.
That second driver, the need to free up capital for infrastructure, is the piece that is widely ignored. It is also the one that does not apply to agencies. Big tech is reallocating billions. You are paying monthly subscriptions.
If you copy their behavior without their constraints, you will create problems inside your business.
This is a common mistake. Operators watch larger companies and copy their behavior without understanding the constraints behind those decisions. Different businesses operate under different constraints, which leads to different decisions.
As we’ve grown multiple agencies toward $10M, we’ve built a different system around AI.
We are still hiring and building teams, but we are changing the slope of that hiring curve. The goal is not to add people at the same rate we did before. The goal is to increase output per person through systems, then layer in hiring as needed.
That is where AI creates leverage.
This newsletter focuses on building the systems that drive predictable growth for agency owners who want to increase revenue, hire more people, and help more customers. Subscribe at businessofagency.com to build systems that produce predictable growth.
We’ve invested in an AI and automation function internally, and it is already changing throughput across the organization. Work that required multiple steps and multiple people is being compressed into fewer steps and fewer people. That does not eliminate the need for talent. It increases what that talent can produce.
Over time, that changes the economics of your system.
You will need fewer people per dollar of revenue than you did in the past. Not because you cut aggressively, but because your systems improved. You will still hire. You will still grow headcount. But revenue should outpace it.
Most agencies miss this. They will react to headlines, assume layoffs are the strategy, and reduce capacity too early. That can improve margins in the short term, but it limits growth because the underlying system does not improve.
The path is building the system first, then letting that system dictate how and when you hire.
If you fix this, everything else gets easier.
~ Erik J. Olson