The org chart that built modern marketing was designed to route information. AI just replaced the router. Agents are the new persona. What comes next is not a reorg. It's a rebuild.
Disclosure: I'm the founder of Goodie AI, a platform for AI search optimization, and the founder of NoGood, a growth marketing agency. The arguments below come from working inside hundreds of marketing organizations, from Series B startups to Fortune 100 brands, as they try to adapt to AI search.
A month ago, I sat in on a review with the CMO of a global consumer brand. Household name. Thousand-person marketing team. They had just wrapped a strategic offsite about AI search and were proud of the outcome: a Manager of GEO with a small team, a real budget, an AEO monitoring platform, all of it slotted under SEO inside performance marketing.
I asked one question. When a buyer asks ChatGPT about your category, who on your team is accountable for the answer? The room hesitated. Then someone said, with confidence, "the SEO team. They own organic search."
I pulled up a live ChatGPT response for one of their highest-revenue product categories. It cited a Reddit thread, a YouTube comparison, a TechCrunch article, and two competitor blog posts. Zero citations to anything the SEO team owned. Their Manager of GEO had no relationship with the social team, no operating cadence with PR, and no read access to the technical logs that would have told her how AI crawlers were interacting with the site. She had budget. She had a tool. She had a metric. She had no authority over any of the surfaces that actually moved it.
This is the structural failure at the heart of how almost every marketing org is approaching AI search. It's not only a tooling problem or a talent problem. Mostly, it's an organizational design problem, and often CMOs are making it worse with every move they think is helping.
Most marketing leaders are treating AI search as a channel
They are wrong, and the cost compounds weekly.
AI search, the constellation of ChatGPT, Google AI Overviews and AI Mode, Perplexity, Gemini, Claude, Copilot, Rufus, and the agentic surfaces emerging behind them, is not a new channel competing with paid, SEO, and social for a slice of the funnel. It is the layer that lives above all of them - where your brand is being assembled, evaluated, and recommended in real time, by agents that will increasingly act on behalf of your customer before your customer ever lands on your site.
Answer Engine Optimization, the discipline required to win, is arguably the most cross-functional mandate marketing organizations have handled for years. AEO is owned. It's earned. It's social. It's technical. All four, simultaneously, fused into a single system whose output is whether an AI model retrieves you, trusts you, and recommends you. And those four disciplines are sitting in four different silos in nearly every large marketing org on earth, often reporting to four different leaders, measured by four different scorecards, meeting maybe once a quarter.
These silos are now a liability. Two truths explain why.
The first: the old marketing org was built to route information through humans. Briefs from strategy to creative, creative to channel teams, channel teams to performance, performance to leadership. AI is now the router. It collapses that middle layer at a speed and scale no headcount plan has priced in.
The second is harder. In an AI and agent-mediated world, you are no longer who you say you are. You are what the internet says about you. Your owned site is one input among dozens. Reviews, Reddit threads, podcast transcripts, affiliate posts, influencer content, support forums, news coverage, your employees' LinkedIn posts, all of it ingested, weighted, and synthesized into a probabilistic representation of your brand that is then served, with confidence, to your next customer. The wider the gap between what your customer experiences and what you claim about yourself, the more aggressively the models will surface that gap to the next buyer.
A marketing org structured for the first reality cannot survive the second.
Why the silos worked, and why they don't anymore
For thirty years, marketing organizations specialized along the seams of the channels they served. SEO optimized for search engines. Social optimized for feeds. PR optimized for journalists. Brand optimized for sentiment. Performance optimized for ROAS. Each function had its own platform, its own playbook, its own metrics, and its own definition of success that could be hit independently of the others.
This worked because the channels were relatively independent. You could run a great social campaign and a mediocre SEO program and the gap would not visibly punish either.
AI search dissolves that independence. When a model assembles an answer about your category, it pulls from your owned content, your structured data, third-party reviews, social discourse, news coverage, forum threads, podcast transcripts, video descriptions, and increasingly, real-time web retrieval. It weighs those sources against each other for consistency, recency, authority, and corroboration. The output is a single synthesized answer in which your brand is either present, partially present, mischaracterized, or absent.
Three things follow, and each one breaks a foundational assumption of how marketing orgs are structured.
One: AEO leans harder on earned and social than on owned. This is the most counterintuitive shift, and the one most CMOs underestimate. For two decades, the gravitational center of digital marketing was your website, the asset you controlled and measured. In AI search, your website is necessary but no longer sufficient. Models disproportionately trust and cite third-party corroboration: independent reviews, editorial coverage, community discussion, expert mentions, structured comparison content. Goodie's research shows social content generating roughly 2.3x more AI citations than owned brand pages. A brand that owns its category narrative everywhere except its homepage will outperform a brand that owns its homepage and nothing else. This inverts the budget logic of the last decade.
Two: technical infrastructure is now a content problem, and content is now a technical problem. Whether a model can retrieve your content depends on schema, semantic structure, crawlability, and machine-readability, historically the domain of technical SEO. Whether it should retrieve your content depends on substance, perspective, and authority, historically the domain of content and brand. These two were already adjacent. Now they are the same job, executed by the same workflow, measured by the same metric. Most orgs still have them on different floors.
Three: social is a citation strategy, not a feed strategy. The job of a social team has always been to drive engagement on platforms: likes, shares, follows, conversion. AI search adds a parallel job: producing social content that is retrievable, attributable, and citation-worthy when an AI is looking for evidence about your category. A LinkedIn post that gets 500 likes and a LinkedIn post that gets cited by ChatGPT in a category-defining query are now equally valuable, and increasingly the latter is more durable. Every piece of social content your brand touches is also training and retrieval data.
The disciplines that have to coordinate to win at AEO, owned content, technical infrastructure, social and community, earned media and PR, are precisely the four disciplines marketing orgs have spent thirty years separating. The silo structure was a feature when channels were independent. It's a tax now that they're fused.
AI is the operating system, not another tool
Most marketing organizations are still treating AI as a productivity tool. A faster way to draft briefs, generate ad variants, summarize research, write first-draft copy. That use case is real. It's also the least interesting thing happening.
The deeper shift: AI is becoming the operating system on which the marketing function runs. The layer between strategy and execution, between teams and channels, between brand and customer. And it is in itself the new audience.
Every marketing function now has a parallel persona to design for. The first is the human customer. The second is the agent acting on behalf of that customer: the AI that researches the category, the assistant that shortlists vendors, the agent that is highly trusted by humans and increasingly executes the purchase. Agents reward structured data, semantic clarity, consistent entity definition, fresh corroboration, and machine-readable claims. Humans reward narrative, emotion, design, and brand. The mature marketing function does both in the same artifact, by the same workflow, without trading off either.
This reframes marketing leadership roles. The work shifts from producing deliverables to orchestrating systems. From writing the brief to designing the agents and skills that produce a thousand briefs consistent with brand. From approving the campaign to ensuring the campaign, the press release, the influencer brief, the help center article, and the schema markup are all telling the same story, because AI will surface every contradiction.
Get the measurement right before you touch the org chart
You cannot manage what you cannot measure, and the measurement stack that got us here will not get us there.
For two decades, marketing measurement has been overwhelmingly deterministic. UTMs, referral data, GA4 events, attribution windows, CRM matchbacks, post-conversion surveys. Every dollar could be traced, with varying degrees of honesty, to a click, a session, a conversion. The stack was imperfect but legible. CFOs could read it.
AI search forces a second, parallel layer that is fundamentally probabilistic. When someone asks an AI a defining question about our category, do we appear? Are we retrieved accurately or wrongly? What is our share of voice across model responses, by query type, by model, by geography? What is the sentiment of how we are described, and is it consistent? Which sources are models pulling from when they mention us?
These metrics do not replace the deterministic stack. They sit alongside it. The mature AEO scorecard fuses both. Probabilistic metrics that prove brand presence in the new discovery layer where revenue increasingly originates, and deterministic metrics that prove revenue impact downstream of that presence.
This is the move that turns AEO from a specialist's job into everyone's job, without the political cost of a full reorg. Each function inside marketing now carries its own retrievability and citation share goals, laddering into a single shared scorecard, not four parallel ones.
Owned owns retrievability of the site and content library or blog. Social owns citation rate of social and UGC. Earned owns the quality and frequency of third-party citations. Technical owns crawlability, performance and agent experience.
Each function's primary goal stays what it has always been. SEO still chases organic rankings. PR still pursues high-impact coverage. Social still builds community. Retrieval rate becomes a secondary goal, measured, reported, compensated against. A PR team told to "do AEO" hears a threat. A PR team told that earned media now has a retrieval function on top of its brand function, and that strong performance against it makes them more strategic, hears an opportunity.
The function that ties them together is brand. Specifically, the unsexy, technical work of entity definition: making sure every surface where your brand appears is telling the same story about what you are, who you serve, and what makes you different. AI is a co-occurrence machine. It learns who you are by watching what concepts appear next to your name, repeatedly, across trusted sources. If those concepts are inconsistent, your entity is fuzzy, your retrievability degrades, and your competitor wins by default.
This is the technical case for brand discipline. It used to be aesthetic. It is now infrastructure.
Structure follows measurement, not the other way around
Most CMOs will be tempted to start the AEO transition with hiring. Don't. Start with the scorecard. Until you can see citation rates by source, retrievability by content type, share of voice by model, and entity consistency by surface, you cannot make defensible org decisions. The measurement layer reveals where the actual gaps are, and those gaps are almost never where the org chart suggests they are.
The aggregate scorecard then needs an owner. Some companies will name a Head of AI Visibility at the CMO's staff level. Some will fold it into an expanded Head of Organic role. Some will create a matrix function that reports both into marketing and into the CMO's office directly. The right answer depends on the company's existing structure, the maturity of its functions, and the political reality of its leadership team.
Structure does not produce alignment. Measurement does. Structure is how you formalize the alignment that the measurement system has already created. A few principles to hold any structure against. Organize around the customer journey and the agent journey in parallel, not around the channels. Treat brand as infrastructure, not as decoration. Shift the senior layer from production to orchestration. CMOs who do not make this shift in their own calendars cannot credibly ask their teams to make it in theirs.
The reality check
None of this is going to be clean. AEO is genuinely new, the measurement stack is still immature, the platforms are changing faster than playbooks can be written, and the temptation to ignore the shift or overcorrect into theater is real on every executive team.
But the underlying logic is settled. CMOs who internalize it now will compound advantage every quarter against CMOs who treat it as a 2027 problem.
The marketing organization was built to control a narrative through channels you owned. That world is closing. The new world rewards organizations that produce consistent, retrievable, credible signal across every surface a machine might learn from. Owned, earned, social, technical, and increasingly conversational. The brands that win in AI search will not be the brands with the loudest owned channels. They will be the brands whose entire internet footprint tells the same story, on the same evidence, in the same voice, to both the human customer and the agent acting for them.
You are no longer who you say you are. You are what the internet says about you. The marketing org that wins the next decade is the one built, deliberately and structurally, to close the gap between those two things.
Everything else is decoration.
