The enterprise AI content scaling strategy is a coordination failure

Conductor's 2026 State of AEO/GEO CMO Investment Report surveyed over 250 executives across 12 industries and asked them what their top content strategy for AI search visibility is. The answer, across every maturity level, was the same: scaling AI content generation. It beat structured data. It beat authoritative long-form guides. It beat original research.

This is the bit where I'm supposed to say "and that's exciting" or "and here's the framework to do it well." I'm not going to. Because the honest read of that finding is that 250 enterprise marketing leaders have collectively agreed to do the one thing the evidence keeps telling them not to do, and they've agreed to do it because they're more afraid of being left behind than they are of being penalised.

That's the loop. And we built it.

What the report actually says, underneath the headline

The Conductor data isn't really a story about AI content. It's a story about what happens when an industry decides a strategy is mandatory before anyone has shown it works.

The people doing the strategy don't believe in the strategy. They're doing it because everyone else is doing it.

Eli Schwartz, quoted in the report, said the leaders he speaks with are "somewhat skeptical about the effectiveness of mass amounts of AI content, but are afraid of being left behind if they don't do this." Read that sentence twice. The people doing the strategy don't believe in the strategy. They're doing it because everyone else is doing it. Aleyda Solis flagged that AI content requires "a personalised editorial and optimisation workflow" to be worth publishing. Lily Ray noted the LinkedIn posts piling up from sites that lost visibility overnight after going aggressive on AI output. Pedro Dias documented that Google started issuing manual actions for scaled content abuse in June 2025 — actual Search Console notifications, citing "aggressive spam techniques."

So the position is: the experts inside the report think it's a bad idea, Google has already started penalising it, and the strategy has nonetheless become the number one priority across the enterprise tier.

That's not a content strategy. That's a coordination failure.

The Mt. AI pattern is now well-documented

Dan Taylor and Glenn Gabe have both written about what Gabe calls the "Mt. AI" effect — the traffic graph shape that's becoming a signature of the failure mode. New AI-generated URLs flood the index, get a freshness boost, traffic climbs sharply for weeks or months, and then a quality assessment threshold kicks in and the whole thing falls off a cliff.

I've been doing this for 18 years and the pattern is identical to every other "exploit a temporary index gap" strategy I've watched in that time. Programmatic doorway pages in 2009. Spun content in 2012. Thin affiliate sites in 2015. The shape of the chart is always the same. A steep climb, a plateau, a cliff. The duration of the climb shortens with each iteration as Google's detection gets faster.

The current iteration looks like it's measured in months, not years.

The strategy isn't failing because AI content is bad. It's failing because scaling without an editorial floor is the same mistake we've already watched fail four or five times in a row.

What Taylor identifies — and this is the bit the industry keeps missing — is that the real problem isn't AI as a writing tool. It's the absence of any content strategy underneath the output. The AI is just an accelerant. If you scale rubbish faster, you have more rubbish faster. The medium isn't the issue. The decision to publish without judgement is.

Why the enterprise tier is the worst place for this strategy

Here's where it gets specifically grim. The enterprise tier is the segment of the market that should, in theory, be the best at this. They have the editorial resources. They have the subject matter experts. They have the brand equity to protect. They have legal review. They have the budgets to do human-guided AI properly.

And yet they're the ones putting it at the top of their priority list as a *scaling* play. Not as a quality play. Not as a "let our experts work faster" play. As a volume play.

The reason, I suspect, is that scaling is the only KPI that survives the trip up the chain. A CMO can show a board "we shipped 4x more content this quarter." They can't show "we shipped fewer pieces but the average citation rate in ChatGPT improved" because nobody has agreed on how to measure that yet. So the metric that's easiest to count wins, and the metric that's easiest to count is volume.

This is how an entire tier of the market ends up running the same strategy the experts inside their own commissioned report are telling them not to run.

The thing nobody is pricing in: the citation portfolio problem

There's a second-order issue here that the scaling debate is completely missing. Recent data on AI citation patterns shows that only about 2.35–2.45% of cited URLs appear across ChatGPT, Perplexity, and Google AI Overviews simultaneously. In other words: optimising for one AI engine is, in most cases, not optimising for the others. The citation surfaces don't overlap.

This means the enterprise "AI content scaling" strategy isn't even a strategy in the portfolio sense. It's a bet that a single content pipeline, pumping out a single style of content, will land citations across three engines that demonstrably don't cite the same things.

If you're going to scale, scale toward something specific. A scaling play that aims at "AI visibility" generically is firing into fog. Most pieces won't get cited anywhere. The pieces that do get cited will get cited on one engine. The ROI maths on enterprise AI content scaling has to account for this and, as far as I can tell from the report, it doesn't.

The honest version of what's going on

Let me give you what I think is actually happening, having watched a lot of these enterprise content decisions get made.

The strategy is a hedge against a fear, not a response to a finding.

The CMOs commissioning these AI content programmes aren't doing it because they've seen the data. They're doing it because their board has asked them what their AI strategy is, and "we're scaling content with AI" is the answer that's legible at board level. It's a defensible position in a meeting. It's also, by every available signal, the wrong one.

This is the part of my job that's depressingly consistent across two decades. The strategies that get implemented at the enterprise tier are not always the strategies that work. They're the strategies that can be summarised in one PowerPoint slide to a non-specialist audience. "We're using AI to publish more" fits on the slide. "We're using AI to help five subject matter experts publish twice as much of the work only they can do, edited by humans against a brand voice guide, structured around questions our customers actually ask, with measurement against citation share rather than volume" does not fit on the slide.

The slide wins. Every time.

What a defensible version of this strategy looks like

If you're going to use AI in content production — and you should, it's a genuinely useful tool when held correctly — the version that doesn't end on the Mt. AI cliff edge has a few non-negotiables.

The output has to clear an editorial floor. Not "we ran it through Grammarly." A floor that says: would this piece exist if AI didn't? If the answer is no, don't publish it. If the answer is yes but better, publish it. The test is whether human judgement is doing the load-bearing work and AI is doing the typing, or whether AI is doing both.

The pipeline has to include first-party expertise. Aleyda Solis's point about "unique brand insights and first-party data" is the actual unlock here. The AI engines cite content that contains things they can't easily synthesise themselves — proprietary data, named expert opinion, original observation. Scaling generic synthesis is scaling the commodity layer that the AI engines are actively trying not to cite.

The measurement has to be citation-led, not volume-led. If you genuinely care about AI search visibility, the metric that matters is what gets cited, on which engines, for which prompts. Not how many URLs you published. Rand Fishkin's recent work with Gumshoe is worth reading on this — AI rank tracking is mostly noise, but visibility share is real, and the distinction tells you what to measure and what to ignore.

And the structural fundamentals have to be in place underneath all of it. There's no point optimising your editorial pipeline if your pages are returning 499s to AI crawlers — Mike King's recent data on the 499 status code as the AI search blind spot shows pages with high failure rates get roughly 18x fewer citation events than stable ones. Eligibility before optimisation. Always.

The counterargument worth taking seriously

The strongest case for scaling AI content goes like this: even if 90% of what you publish gets ignored or eventually penalised, the 10% that lands creates citations and brand mentions that wouldn't otherwise exist, and the unit economics work because AI production is so cheap that the failed 90% is essentially free.

I don't think this is a stupid argument. I think it's a survivable argument for some businesses — specifically, businesses with low brand risk, no editorial reputation to protect, and the technical capacity to detect and remove the failed cohort before it drags down the domain.

For an enterprise with 18 years of accumulated domain authority, a recognisable brand, and a legal team that cares about what's published under the masthead, that calculus is much worse. The downside of a manual action on your primary domain is not "the failed cohort gets removed." It's a sitewide trust hit that takes a year to recover from. The risk-adjusted maths flips.

Which means the strategy that might work for a thin affiliate site is the same strategy that creates existential risk for the enterprise tier currently rushing to adopt it.

The honest limits

I should be clear about what I'm not saying. I'm not saying AI is bad for content. I use Claude every day. The pieces on this site are AI-assisted and I'm fine telling you that, because the editorial guidance, the positions, the source briefs, and the standards are mine. The work happens before and after the typing.

I'm also not saying enterprises shouldn't be experimenting. They should. The pace of change in AI search makes experimentation a requirement, not an option. The mistake isn't experimenting. The mistake is naming the experiment as the top strategic priority before the evidence supports it, and committing volume to a play that the people designing it don't believe in.

And I might be wrong about the timing. It's possible that the Mt. AI cliff edge that's hit a number of sites in 2025 is a transient phenomenon that gets engineered around. It's possible Google's detection capabilities plateau. I don't think it's likely — the history of search quality systems is that they always get sharper, never duller — but I'd be lying if I said I knew the exact trajectory.

Where this lands

The Conductor report is going to get cited a lot over the next six months as evidence that AI content scaling is "what enterprises are doing." It will be used in sales decks by tool vendors. It will be cited in agency pitches. It will be paraphrased into board presentations as proof that the strategy is mainstream.

What it actually shows is that 250 enterprise leaders have agreed on a strategy that the people they pay to advise them are openly worried about, that Google has started penalising, that has a documented failure pattern with a name, and that ignores the fundamental citation-portfolio problem sitting underneath the entire AI search measurement question.

The race-to-the-bottom on AI content scaling isn't going to be won by whoever publishes the most. It's going to be survived by whoever publishes the least junk. Those aren't the same metric. They're not even the same direction.

If you're a marketing leader staring at a slide deck that recommends scaling AI content as your number one AI search priority, the question to ask is not "how do we scale faster than competitors." It's "what's the editorial floor under this, and who's accountable when it fails?" If nobody can answer that confidently, you don't have a content strategy. You have a coordination problem dressed as one.