A marketing CEO just cut his staff in half because clients started using AI themselves, and he's calling it the best thing that could have happened to his business.
The Summary
- Alex Cohen's 17-year-old marketing agency lost clients and revenue after ChatGPT launched, forcing him to cut his already small team by 50%
- Clients didn't fire him because his work was bad, they pushed for lower fees or left because they could do content creation themselves with AI
- Cohen responded by pivoting his business to sell AI integration services to in-house marketing teams, betting the future is helping companies build their own AI workflows instead of doing the work for them
The Signal
This is what the middle of a labor transition looks like. Cohen saw ChatGPT in November 2022, spent Christmas testing it, and told his team on day one of 2023 they needed to figure out how to use AI. He was early. He was proactive. It didn't matter.
By summer 2023, a repeat client asked to reduce the scope of content creation services. Another asked for lower prices on the same work. The requests weren't about quality. They were about necessity. When clients can generate acceptable marketing copy in seconds, paying an agency becomes harder to justify.
Here's what makes this story matter: Cohen isn't bitter. He calls AI a "double-edged sword" but remains more excited than worried. He cut staff and launched a new service line teaching companies how to integrate AI into their own workflows. He's betting that if clients won't pay him to create content, they'll pay him to teach them how to create it themselves at scale.
This is the quiet part most founders won't say out loud: AI isn't just coming for repetitive jobs at big companies. It's coming for skilled service businesses run by smart people who saw it coming and still couldn't stop it. The difference between Cohen and the agencies that will close in the next 18 months is that he's treating this like a product market fit problem, not a temporary downturn.
The Implication
If you run a service business where the output can be generated by an LLM, you have maybe 12 months to figure out what you sell when clients don't need the service anymore. Cohen's answer is implementation expertise. Yours might be different. But pretending the work will come back at the old price is not a strategy.
Sources: Business Insider Tech | Business Insider Tech