On Marketing
Marketing Is About to Look Very Different
The story of one client, one AI system, and what it taught me about where the marketing industry is actually going. A note from inside the shift, not a prediction from outside it.
One client, one AI system, and what the build revealed about where the whole industry is going.
I had to part ways with a client recently.
The relationship had been good. The work was working. But circumstances changed, and eventually I had to make the call. So I did.
What I didn't expect was what she said to me on that call. Not anger. Not defensiveness. Just: How am I going to do this without you?
And in that moment I made a decision I didn't plan on making. I told her I was going to teach her. I was going to build her a system so she could run this herself.
The system we built together is what this essay is about. Because what I saw inside that build is what the next three years of marketing is going to look like for everyone.
What the system actually does
The system runs her Meta ads. She still creates her own ad assets, the visuals, the videos, the founder content. But everything else is automated.
Every day, the system pulls the last seven days of campaign performance and compares it against the previous seven days. It looks at ROAS, cost per purchase, click-through rate, frequency, and a dozen other metrics. It runs logic against those numbers, actual decision-making logic instead of just reporting, and sends her a text message every morning summarizing what's happening and what to do about it.
If an ad is starting to fatigue, the system catches it within a couple of days and tells her she needs to refresh the creative. If a headline is dramatically outperforming the others, the system flags it and suggests building more variations around that angle. If cost per purchase starts drifting in the wrong direction, the system notifies her with specific optimization suggestions rooted in her numbers, not generic playbook advice.
It pulls best-performing headlines. Best-performing primary text. Best-performing creative angles. It tells her what's winning, what's dying, and what to test next.
You used to have to pay a person to do all of this. Now her phone tells her.
What this means for the marketing industry
What I built her is not unique. Versions of this system are being built right now by smart people across the industry. Some are buying off-the-shelf tools that approximate it. Some are building custom workflows the way we did. The specifics vary. The trajectory doesn't.
The numbers already show up in the reporting. Forrester's 2026 predictions forecast a 15% reduction in agency jobs this year, after an 8% cut in 2025.[1] Meta plans to fully automate ad creation, targeting, budget, and optimization by the end of 2026, a system where you feed in a product URL and a budget and the AI builds the rest.[2] And 60% of senior US marketers say they're spending less on agencies as a direct result of AI.[3]
Over the next one to three years, that trajectory continues along three predictable lines.
Traditional marketing agencies will shrink. Not all of them, and not all at once. But the agencies whose value proposition was "we have a team that can execute" are about to discover that one founder with the right AI system can execute at a level that used to require five people. The economics don't hold anymore.
Specialist execution agencies will be hit first. Pure white-label SEO agencies. Pure graphic design shops. Single-platform paid media agencies whose only differentiator was that they did the work the client didn't have time for. That work itself is now something a well-built AI system can do, often faster, cheaper, and more precisely than the agencies were doing it. Industry analysts are calling it the labor-based model collapse, and it's not a slow one.[4]
What will survive is the strategic and human layer. The agencies that survive will be the ones doing the work AI can't do. Brand strategy. Founder psychology. Pattern recognition across complex business situations. The kind of listening AI doesn't yet do well. Presence and judgment.
Personal brands, in particular, are going to be one of the last things AI replaces. Because what makes a personal brand work isn't the content production. AI can do that now. What makes it work is the soul underneath. The specific voice. The lived experience. The reason the founder is the one saying this and not someone else.
AI can't replicate that. It can only support it.
What founders are getting wrong right now
The mistake I see most often is founders underusing AI in their own businesses.
Most of the founders I talk to are still treating AI like a slightly better Google. They're using it to write social captions and draft emails. That's fine, but it's a tiny fraction of what's actually possible.
What founders are missing is that AI can build entire systems for them. Not a single caption, but a content engine that runs on its own. Not one ad headline, but a system that monitors every ad in every campaign and tells them what to do every morning. Not a website draft, but the whole site, all the copy, the SEO setup, and the deployment, in a day.
The moment matters, too. For the first time in history, more than one in three new companies in 2026 was founded by a single person.[5] A generation of solo founders is walking into a landscape where AI can do most of what an agency used to charge them for. If you're a founder running a service-based business, a personal brand, a small e-commerce shop, or really any business under a certain size, you don't need an agency anymore. You need a system. And the right person to help you build the system, once, so you can run your own marketing for the rest of your business's life.
That's a different kind of partnership than what most agencies sell. It's also where I think the value in a strategy partnership is actually going.
AI systems aren't perfect
What nobody building these AI workflows wants to admit: the systems mess up.
They pull wrong numbers sometimes. They misread a dashboard and report the metric off. They draw the wrong conclusion from a pattern. They suggest the wrong optimization because they read the data without understanding the context.
This is real, and it's not going away anytime soon. Recent hallucination research shows that even after a roughly 95% reduction in factual error rates from 2024 to 2026, top models still hallucinate on multi-turn, cited tasks at around 30% in the hardest benchmarks.[6] The systems are getting better fast. They're not infallible. And most founders don't have the marketing expertise to know when the AI is wrong.
That's the part of the equation I think people miss when they talk about AI replacing marketers entirely. You can build a beautiful automated system, and it can still be confidently incorrect about your ad performance for three weeks before anyone notices, unless someone in the loop actually understands the methods, the metrics, the platforms, and can catch the error.
Which is to say: the human work isn't going away. It's shifting. Strategy and oversight and judgment matter more now than when humans were doing the execution, because the execution is happening fast enough that errors compound before you see them.
The right pairing is AI doing the heavy lifting with a human who actually knows what they're looking at standing watch. That's how the systems work well. Without that, you're just trusting a robot with your business.
What you bring that AI doesn't
Here's the part nobody talks about enough.
AI can run your ads. It can write your captions. It can analyze your numbers. It can build your systems. It can produce content faster than you ever could.
But it can't be you.
It can't bring your specific lived experience to the work. It can't know why you started the business. It can't carry the mission you carry. It can't show up on camera as a person other people can trust. It can't speak the Truth that's already inside you, because that Truth isn't in any training data. It's in you.
The future of marketing is going to be founders who understand this distinction clearly. Founders who use AI for the things AI does better than humans, the execution, iteration, optimization, and production, and then show up themselves for the things only they can do. The voice. The vision. The trust. The relationship. The soul.
That's the shape of what's coming. The agencies that survive will be the ones that help founders make this distinction and build the systems that support it. The founders who thrive will be the ones who stop trying to be everything and start being the one thing AI can't replace.
What I'm doing about it
The studio is built for this future. I run on a layered system of AI tools that handle most of the execution layer. I spend my time on the strategy, the alignment, the listening, the human work. For clients who want to go further, I build them their own systems, like the one I built for the client whose story started this essay.
Some of those clients keep me on as their strategist forever, with AI running the day-to-day. Some reduce their reliance on outside help entirely. Either way works. The point is that the old model, paying an agency a lot of money to do work that a founder could now do themselves with the right system, is breaking down in real time. The industry analysts are seeing it play out in the numbers.
If you're a founder, the question isn't whether AI is going to change your marketing. It is. The question is whether you're going to let it shape you or shape it yourself.
The work is figuring out what you bring that AI doesn't. And then building everything else around that.
References
- [1]Predictions 2026: Marketing Agencies Resign Their Agency · Forrester
- [2]How Meta's AI push is changing ad creation · Marketing Brew
- [3]AI is squeezing marketing agencies from both sides · Search Engine Land
- [4]The Agency Reset 2026: AI, Procurement, and the Collapse of the Labour-Based Model · Piscari
- [5]Solo Founder Marketing 2026 ‚Äî From Zero to First Customers · Lishchuk
- [6]AI Hallucination Statistics 2026 · Suprmind (research aggregator)
— Chloe
Founder · Miami, FL
If this resonated, let's talk about your work.