Marketing is broken. What’s next?

Marketers, you're right to be worried about AI's impact on your job. You should probably be more worried than you are.

You've heard the take. AI will automate the execution, but strategy will always need a human. The entry-level jobs go, the senior ones stay. The content writers are at risk, the CMOs are fine. You might even believe it, because it's reassuring.

It's not wrong, but it's not right either.

Here's what I actually know: I could take ten years of my frameworks - the way I think about ICPs, positioning, messaging architecture, channel strategy - upload them, and build an agentic version of myself that runs a competent marketing operation of ~6 agents and an operator agent. Not perfect. Maybe 80%. But 80% at a fraction of the cost and none of the salary, and it ships on Sunday at 2am without complaint. That's possible right now. Not in five years. Now.

So when someone tells you strategy is safe, ask them: safe from what, exactly? Safe from a version of you that runs 24 hours a day and never needs a performance review?

How We got here

Modern marketing has moved through distinct phases, and understanding them matters for understanding what comes next.

1. It started as push. The Mad Men era. Marketers and advertisers owned the channels (television, print, radio), so you told people what was good and they more or less believed you. Distribution was the moat. You controlled what people saw.

2. Then it became respond. Inbound, content marketing, SEO. The consumer got the internet and suddenly they were searching, not waiting. The job shifted to being findable when someone came looking. You stopped shouting and started answering. You got data and you responded, you saw what worked and you did more of it.

3. Then it became predict. Programmatic, algorithmic, social signals. You stopped waiting for people to come looking and started anticipating where they'd be, what they'd want, before they knew it themselves. The job became data analytics: reading signals, modeling behavior, showing up in the right place at the right moment.

And right now, most of us think we're moving into AI-optimized marketing. AI content. AIO. AI visibility. Making sure your brand shows up when someone asks their AI assistant a question instead of typing it into Google. Getting into the model's training data. Becoming the answer the agent gives.

Here's the uncomfortable truth about that: it's the same game on a different surface.

Predict, but for machines instead of humans. Optimize, but for the algorithm's algorithm instead of the algorithm. We've been so focused on adapting the playbook that we haven't asked whether the playbook still works at all.

AI collapses marketing’s execution layers, but it also collapses marketing as a function itself.

What AI Actually Breaks

When AI intermediates enough decisions, something structural shifts. Not just “how marketing is executed,” but in whether it works.

Think about what marketing has always depended on: distribution choke points. Attention was scarce, channels were limited, and if you could get in front of someone at the right moment with the right message, you could shape what they thought and bought. Every tactic we've ever used - ads, content, SEO, social, influencer, email - is a variation on "reach the human before they make up their mind."

AI disrupts the premise, because everyone has a perfectly personalized personal assistant in their pocket. When someone asks an assistant "what's the best CRM for a 50-person company" and gets three recommendations, the entire funnel that existed between that person and that decision has been collapsed. The awareness campaign, the comparison blog post, the retargeting ad, the nurture sequence…gone. Bypassed. The agent made the shortlist. If you're not on it, your marketing might as well not exist.

And agents don't respond to marketing the way humans do. They don't have feelings about your brand. They don't remember the clever campaign. They don’t care about cool.

They evaluate on structured signals: performance data, verified reviews, compatibility, reliability metrics, ecosystem integration. This is what I'd call the shift from a persuasion economy to a verification economy. You don't win by making people feel something. You win by being the option that the system trusts.

And B2B is not safe. This feels like the kind of shift that would be uniquely critical for consumer brands, but that’s a fallacy. B2B purchasing has historically been protected by relationship complexity, long sales cycles, and the political dynamics of organizational buying. That protection is eroding fast. High-pressure execs who have had the sinking feeling that their team isn’t really working all that hard, that their MarTech stack isn’t doing what it could, now have access to something that says, “You’re absolutely right,” and then gives them the answer. It’ll do more: evaluate alternatives, model switching costs, and produce a shortlist before the QBR even happens. The RFP process, vendor selection, renewal decisions: these are all pattern-recognition problems that AI handles well. The relationships that protected B2B incumbents were always partly a substitute for better information. AI is the better information. That’s why AI might actually be most disruptive for B2B marketers, who have long been somewhat insulated from the volatility of consumer behavior. That insulation is gone.

Meanwhile, for the decisions that are still human - and many are, for now - there's a different problem. The volume of content has exploded. AI generates infinite blog posts, infinite ads, infinite personalized emails, infinite social content. The channel is full of noise that sounds exactly like marketing. People are overwhelmed. Attention isn't just scarce; it's defended. “Oversaturation” doesn’t even cover it. People have developed immunity.

Scalability, as we've understood it, is breaking. The spray-and-pray model is broken. The content flywheel is broken. The funnel is broken. And yet most marketing budgets and strategies are still organized around the assumption that these things work. They don’t (or won’t). That’s why AI is really coming for you.

AI is a perfect scale machine. It does content production, optimization, reach, and personalization at volume faster and cheaper than any human operation ever could. And scale - the ability to reach the maximum number of people for the minimum cost - is what modern marketing became. Somewhere in building the martech stack, the funnel, the content flywheel, the programmatic machine, the underlying purpose got inverted. Marketing stopped being "understand people well enough to be genuinely useful to them" and became a scale delivery system. The metric ate the mission.

But marketing was never supposed to be a scale operation. Originally it was about understanding people - what they needed, what they believed, how to speak to them honestly (or dishonestly) enough that they'd choose you. The craft was relational. A good salesman knew his customers. Scale came later, and it was so seductive that it swallowed everything.

Most marketers believe that their access to data, their ability to create content, their unique strategic POV holds intrinsic value. It doesn’t. Not anymore.

AI ends the detour. If your value is scale, you just became redundant. The question is what was underneath it.

The Real Shift: Distribution to Trust

If I had to reduce everything I've been thinking about to one sentence, it's this:

Distribution architecture is dead as a competitive moat. Trust architecture replaces it.

Distribution architecture is the old game: own the channel, control the reach, spend enough to stay top of mind. It worked when attention was capturable and people made decisions based on familiarity.

And for a long time, trust architecture was just a lazy version of the same thing. When there were only a handful of channels and none of them were democratized, you could own a television network or a national magazine and call it trust. That wasn't trust. It was captive attention. The channels were scarce enough that you didn't have to earn belief, you just had to show up consistently in a place people had no choice but to look. That's gone. Channels are now infinite and democratized. Captive attention is gone. Which means you actually have to earn it.

Trust architecture is the deliberate construction of the signals, relationships, systems, and proof that make your brand the one someone - or something - turns to when it matters. It's not about being everywhere. It's about being believed somewhere.

What trust architecture actually looks like varies enormously by brand, category, and consumer. There's no single playbook. But there are two distinct audiences to build trust with now, and they respond to completely different signals.

Humans respond to provenance, community, social consequence, and consistency over time. In a world saturated with synthetic content, demonstrably human origin becomes legible as different in kind, not just quality. The essay someone clearly read. The recommendation from a person with skin in the game. The brand with a point of view consistent for years, not generated fresh each morning. This is one of the most durable trust signals available right now, and yet most brands are actively undermining it by automating everything they can.

Meanwhile, brand marketers beat the drum of “authenticity.” That instinct is right, but authenticity as a value won't be a bulwark against the onslaught AI. It has to be structural.

Agents respond to performance data, structured proof, and ecosystem integration. For the decisions that are increasingly agent-mediated, the trust signal isn't the story you tell about yourself. It's the proof you've accumulated, like return rates, customer outcomes, third-party validation, ecosystem integration, reliability over time. Consumer brands should have been tracking customer fallout all along, but volume of sales and traditional visibility have masked all manner of ills. That cover is gone.

The strategic question - and it's a real one - is knowing which audience you're building for, when, and whether you can hold both without one corrupting the other. Optimizing hard for agent legibility can actively undermine human trust signals. It starts to feel corporate, synthetic, like something assembled rather than meant.

Community is infrastructure, not tactic. I want to be precise about what I mean here. Community isn't a loyalty program or a Discord server you stand up because a consultant told you to. It's the condition where your customers are in relationship with each other around something you enable or represent. That's rare and hard to fake, which is exactly why it works as trust architecture. It also doesn't scale the way distribution does — and that's the point. The things that don't scale are becoming the things that matter. Kiss scalability, and predictable growth-at-all-costs, goodbye.

Trust requires the possibility of consequence. The reason a recommendation from a fellow human is more trustworthy than an agent's isn't that humans are more accurate. It's that humans have something at stake. Trust implies the possibility of betrayal. A friend who recommends a bad restaurant loses a little credibility. An AI agent loses nothing. Brands and individuals who put their reputation on the line, who are visibly accountable for what they recommend and associate with, carry a weight that optimized content never can. And eventually, even AI systems will begin to bake this into their evaluations.

Are You Actually Worth Trusting?

Before you audit your trust architecture, ask a harder question: are you actually worth trusting? And be honest about whether the answer requires you to stay smaller than you planned.

The brands I actually trust were built by practitioners who value the same things I do. I trust my gardening tools (a Kent & Stowe spade and pitchfork) because they were built by people who expected you to use them weekly for years. They last. The materials feel right in your hands. That's not a marketing decision. It's a values decision that preceded any marketing, expressed through the object itself. I trust these products because they were built by a company that holds the same values I do. Contrast that with the cheap hardware store version that works fine until it breaks and gets replaced. Both are rational business models. Only one builds trust.

That distinction matters enormously right now, because it points to where trust architecture is actually buildable…and where it isn't.

Physical reality is genuinely resistant to the optimization loop. A spade can't be enshittified by a software update. Durability, material quality, and feel are real and can't be faked at scale. An agent can recommend a cheaper alternative, but it can't make the cheaper alternative last twenty years. The grounding that people are increasingly craving - the desire for things that feel worth having, that connect to physical reality, that don't need to be replaced - is not sentimental. It's structural. It's a category of human experience that optimization cannot fully reach.

Identity-based choices still need storytelling. Content strategy, editorial voice, and brand storytelling matter when you’re selling something that relates to a person’s identity, community, or self-perception. And this is where “authenticity” can still add value, because these are not explicitly rational decisions. It’s trust, but again, not really scalable.

Digital is different, and we should be honest about it. I don't really trust any of my digital tools anymore. I'll change if something better comes along, and the switching cost is approaching zero. That's a rational response to an entire category that has optimized for retention and engagement rather than actually being worth trusting. In digital, the enshittification is real and people feel it even when they can't name it. If you're building a digital product, the trust question is harder and the answer is less stable.

The VC question buried in all of this is pointed. The brands that can build genuine trust tend to be small, focused, and built by people who use what they make. Scale corrupts the signal. A gardening tool company that grows large enough to need a procurement department and a quarterly earnings call is a different company than the one that made the spade. This is an uncomfortable thing to say, but it's true, and it raises real questions about what growth models are actually compatible with trust architecture.

A Bifurcation worth naming

AI plus automation plus globalization = commoditization. We can have everything we’ve ever needed, for less than it’s ever costed. That will most likely accelerate. When human signal becomes the premium trust marker, when craft, provenance, genuine community, and accountable recommendation become what people seek out, who has access to those things?

The honest answer is that genuinely human experiences, products, and networks will probably be premium first. The mass market gets agent-curated, algorithmically assembled, optimized-for-engagement everything. The people with resources get provenance, craft, and trusted human networks. We're already seeing it: quiet luxury, the return of analog, the explosion of niche subcultures as identity signals. These aren't random aesthetic trends. They're responses to a world where everything can be generated and nothing feels like it means anything.

The dystopian sci-fi version of this future shows everyone dressed the same, living the same, the masses flattened by optimization, the classes bifurcated. I don't think that's inevitable. Vinyl was an audiophile luxury and then it was everywhere. Farmers markets were bougie and then they weren't. Human-signal goods and communities have historically diffused downward once the premium market establishes their value. We may see that cycle continue.

For brands, this means something specific: where you position yourself in relation to the human-signal premium is a strategic choice you're making right now, whether you're conscious of it or not.

A Chorus of Whatabouts

I can hear the objections:

"What about brands that can't afford to build trust architecture? This sounds like advice for companies with resources."

Fair. Not every brand can build a genuine community or invest in the kind of craft that makes a Kent & Stowe spade. But the affordability question cuts both ways. The brands that can't afford trust architecture also can't afford to keep doing what they're doing, because the channels are breaking and the content flywheel is dying regardless of whether you have a better alternative ready.

The smallest brands may actually have an advantage here: they're closer to their customers by necessity, they haven't yet built the abstraction layers that large brands use to avoid actually knowing who they serve, and genuine community is more achievable at 500 customers than at 500,000. The cost isn't money. It's the willingness to stay specific and resist the pressure to scale before you've earned it.

"What about large brands that already have trust? Nike, Apple, Patagonia, aren't they fine?"

They earned that trust in a window that is closing. They owned channels that no longer function as chokepoints. They had cultural authority in a pre-fragmented media landscape that new entrants cannot replicate. And the data suggests that trust isn't holding as well as it looks: the world's 50 largest CPG brands grew just 1.2% in the first half of 2024, while insurgent brands captured roughly 40% of overall consumer products growth. 71% of global consumers said in 2024 that they trust companies less than they did a year ago. Large brand trust is eroding slowly, then - if the historical pattern holds - quickly. What those brands have is time and margin. Whether they use it to actually rebuild trust architecture or just defend distribution share will determine which ones survive the next decade.

"Isn't this just another 'marketing is dead' cycle? People have been saying this since the death of TV advertising."

Yes, and those predictions were usually wrong about timing and scope. But they weren't always wrong about direction. TV advertising did decline. Organic social did collapse. SEO did get disrupted. Each of those was a channel shift. This is different because AI doesn't disrupt a channel, it disrupts the underlying mechanism that all channels depended on: a human making a decision that marketing could influence. When the decision-maker is an agent optimizing on structured data, the entire persuasion stack loses its footing at once, not just one platform at a time. That's not a cycle. That's a structural change to the substrate.

"What about regulated industries like healthcare, finance, legal? Aren't they protected by compliance requirements?"

Partially, and temporarily. Compliance requirements do create friction that slows agent-mediated purchasing. But regulation follows behavior, it doesn't lead it. The agent-mediated purchasing patterns being established right now in low-stakes categories will migrate upward as trust in those systems builds and as regulatory frameworks catch up. Healthcare and financial services have already seen significant AI-driven disruption in information discovery, and the purchasing layer is next. If anything, highly regulated industries should be building trust architecture more urgently, not less, because the window before disruption hits is still open but it won't stay open.

"What about markets outside the US? Does this argument hold globally?"

The structural argument that AI disrupts distribution choke points holds wherever AI agents become prevalent, which is happening at different speeds in different markets. The trust architecture response looks different by culture. In relationship-first markets, the human provenance signal was already primary; the shift there is less about reclaiming trust and more about protecting existing relationship structures from being undercut by cheaper automated alternatives. In markets with different platform dominance, the agent layer will look different. But the underlying dynamic (scale breaking as a moat, trust becoming structural) is not uniquely American. It follows wherever the optimization loop goes.

What I'm Watching

I want to be honest: I don't have a clean resolution to offer. The trust architecture frame feels right to me. The components feel real. But the specific shape of what marketing looks like on the other side of this. I'm still working it out.

Here's what I'm paying attention to.

How AI agents develop their own trust signals. The verification economy isn't fully formed yet. Right now agents rely heavily on the human-generated signals that already exist, like reviews, ratings, structured data. But those were built for human cognition. At some point agents will develop mechanisms to evaluate each other: performance telemetry, cryptographic credentials, ecosystem reputation networks. When that happens, the trust game changes again. The brands that are building genuine performance data and structural proof now will be better positioned for that world than the ones chasing AI visibility through content.

Whether "totally personalized at scale" is actually possible, or advisable. The promise of AI is that every consumer gets a perfectly tailored experience. But totally personalized at scale may produce a kind of uncanny valley effect: the experience that is optimized for you but doesn't feel like it came from anywhere. Whether consumers learn to distrust that, or whether they just stop caring, is one of the more important open questions for brand building.

What identity actually means when you can present as anything. This is the one I find myself coming back to. AI collapses the cost of presentation. You can look like, sound like, be associated with almost anything. When presentation is infinite, identity shifts to something else: what you consistently do, what you build, what you stand behind over time. For brands, this suggests that consistency and genuine commitment become more valuable, not less. The brands that have always stood for something specific, and paid the costs of that commitment, may find themselves to be even more resilient than they thought.

And what happens when trust isn't even the variable anymore. If AI evaporates switching costs - and it will, because the friction of trying something new drops toward zero when an agent can migrate your data, learn your preferences, and onboard you to a competitor in minutes - then even earned trust doesn't create lock-in. You can be genuinely beloved and still lose the customer the moment something marginally better exists. And if the decisions are increasingly made below the surface, agent to agent, optimization to optimization, there's no human moment to intervene in. No raised hand. No consideration set. Just a continuous background process reallocating spend toward whatever performs better this week.

For commodity categories, for digital tools, for anything where switching costs are low and performance is measurable, that onward grind probably wins. Those markets become fully automated, fully optimized, fully inhuman. And the job of marketing in those categories really does become just visibility and performance data. Which is mostly an engineering problem.

The most important strategic question, maybe the only one that matters right now, is whether trust architecture can protect you.

If you're in a category where trust is buildable and durability is real, build trust architecture deliberately and be willing to stay smaller than the growth model demands. If you're in a category where switching costs are gone and agents make the calls, stop pretending marketing solves it. That's a product and infrastructure problem.

What This Means for Marketers

Marketers, I want you to be able to do the job you love. I want for that to still exist for you. I want for it to exist for me.

But I think you need to be prepared for your options to change. The middle layer of marketing - the execution, the optimization, the strategy built on frameworks that can be uploaded and replicated - that layer is compressing. Not eventually. Now.

What remains is harder to name and harder to hire for: the judgment about what's worth building, the relationships that can't be automated, the conviction to stand for something specific over time and pay the real cost of that commitment. Whether that still gets called "marketing" or dissolves into something else, I genuinely don't know.

My fellow growth marketers have long been positioned as a bridge between customer, product, company and competitors, and this “spider in the web” role will likely be valuable for relatively longer. But if you like doing the work, you’ll be looking for the types of companies that have a value system that aligns with paying you to do it, versus bringing an AI architect in.

What I do know is that the people who will navigate this best are the ones asking the hard questions now, not after the transition has already happened.

And only you can answer what you want your future to look like. Luckily, we’re human and we get that chance.

What Comes Next

The AI-powered optimization loop isn't just capturing attention, it's colonizing judgment. And when you lose judgment, you lose something more basic: the ability to feel what is right.

This is why what comes next is not just a shift in marketing. It’s a shift in human behavior.

As systems take over decision-making, people don’t just need better answers. They need to recover their own connection. When you outsource enough decisions, you start to lose the signal from your own body that tells you what you actually want. Humans today don’t need more orientation; we have more direction than ever.

What that looks like is grounding.

The craving for things that feel real isn't just sentimentality or nostalgia. It's a recalibration. And a brand that understands this isn't just building trust architecture. It's offering something rarer: a reason to be present. Brands that figure out how to be genuinely worth a person's presence, not just their attention, are building something that no optimization loop can fully reach.