Something Inc.LoginSchedule a free consultation
CONTENT

Your Content Didn't Get Worse. The Room Got Louder.

Marketers followed the AI content playbook to the letter this year: more output, faster. New survey and ranking data say performance fell anyway, and the reason was never quality. It was volume.

CONTENT MARKETINGDATA TEARDOWNAI CONTENT

You spent this year doing exactly what every conference talk told you to do. You wired AI into the content pipeline, doubled the output, hit publish more than you ever have. And the number everyone was chasing didn't move. Or it moved the wrong way. You did not do this wrong. The room got louder, and almost nobody adjusted their voice to match it.

KEY TAKEAWAYThe overwhelming majority of marketing teams increased AI content output this year, and only 6% say it meaningfully improved performance. The failure mode showing up in the data is not content quality. It is content saturation: more competent, AI-assisted material chasing the same finite set of queries and inboxes, which quietly cuts the expected attention any single piece can earn.

Everyone Followed the Same AI Content Playbook

This was not a rogue minority chasing a shortcut. It was the default move, adopted at near-universal scale. In 10Fold's AI-First, Buyer-Ready report, 91% of marketing teams say they plan to increase content output again this year, and 46% expect to produce three to five times more than they did before. A separate Canto and Ascend2 survey puts a number on how far that shift has already gone: 75% of content professionals report that AI increased their production volume, and only 4% say they are not using AI for content work at all.

The logic behind the bet was reasonable on paper. More surface area should mean more chances to rank, more chances to get cited, more chances to catch a query nobody had covered yet. Teams we work with inside our content marketing engagements made this exact calculation heading into the year: if one well-built page earns a certain amount of attention, five pages should earn more, even accounting for some quality drag from the faster pace.

That math assumes the rest of the market holds still while you move. It did not. Every competitor reading the same conference circuit and the same vendor case studies made the identical bet at roughly the identical moment, which is how an entire industry can independently arrive at the same overcorrection without anyone actually coordinating it.

The Performance Gap Nobody Predicted

31.4%
of marketers name organic search and SEO as the area that declined the most this year
91%
of marketing teams plan to publish even more content this year
6%
of B2B marketers say AI meaningfully improved content performance
23%
average ranking gap for unedited AI content versus human-written work, and still widening

CoSchedule surveyed 911 marketing professionals in December 2025 for its After the AI Shift report, asking a direct question: since adopting AI at scale, what got worse. The top answer was organic search and SEO performance, named by 31.4% of respondents, ahead of general website traffic drop-offs at 21.7% and email marketing losses at 21.4%. Only 26.1% of the same group still call SEO their highest-ROI channel, down from a much larger share in prior years. CoSchedule's full write-up breaks the decline out by channel, and organic search sits alone at the top of the list, ahead of every paid or owned channel the survey asked about.

The number that stings more sits in a separate survey. MarketingProfs and Storyblok asked B2B marketers in August 2025 whether AI adoption had meaningfully improved their content performance. Just 6% said yes. Not 6% saying AI made things worse — 6% saying it clearly helped. The other 94% got a faster pipeline and a flat or falling result, which is precisely the shape you would expect if the binding constraint was never how fast you could write.

AI Content Saturation Is the Actual Mechanism

CoSchedule's own survey did not stop at the headline number. It asked marketers to name the mechanism behind the decline, and the top answer was content saturation: more AI-generated material competing for the same queries and the same inbox space, which reduces the visibility available to any individual piece even when that piece is competently made. The other two named mechanisms were a discovery shift toward AI summaries and zero-click environments that mediate content without sending a click back to the source, and a measurement gap where content increasingly shapes awareness without leaving an attribution trail.

AI-assisted, mixed with human editing71.7%
Pure human, no AI involvement25.8%
Pure AI, unedited2.5%

Composition of 900,000 newly published English-language web pages, by AI involvement (Ahrefs, Ryan Law and Xibeijia Guan, April 2025).

That composition chart is the saturation problem in one picture. Only 2.5% of new pages are raw, unedited AI output — the caricature most people still argue about when they debate whether "AI content" is good or bad. The real shift is the 71.7% in the middle: competently AI-assisted, human-edited pages that read fine individually and, in aggregate, have multiplied the number of credible competitors for every query worth ranking for. Saturation does not look like spam. It looks like a normal, well-edited article, published next to several hundred thousand others that also look fine.

Layer a shrinking click pool on top of a larger competitor set and the arithmetic gets worse from both directions. Seer Interactive tracked click-through rate on the same query set from June 2024 to September 2025 and found it falling everywhere, faster where an AI Overview appeared: queries without an AI Overview went from a 2.74% average CTR to 1.62%, and queries with one went from 1.76% to 1.01%. More pages are chasing a click pool that got smaller over the same period. Saturation and shrinkage are the same story told from two sides.

Where AI Content Saturation Hits Hardest

The saturation effect is not evenly spread across every keyword you publish for, and that unevenness is the most useful finding in the data. Digital Applied tracked 4,200 articles across 140 domains — 1,400 pure AI, 1,400 AI-assisted, 1,400 fully human-written — weekly, across search positions 1 through 50, for sixteen months. Pure AI content ranked 23% lower than human-written work on average. But the size of that gap depended almost entirely on how contested the query already was.

KEYWORD DIFFICULTY TIERPURE AI RANKING GAP VS. HUMAN-WRITTEN
Low (KD 0–25)8%
Medium (KD 26–50)22%
High (KD 51+)41%

At low difficulty, pure AI content was nearly competitive: an 8% gap most teams could live with. At high difficulty, the gap was 41% — the exact territory where saturation is worst, because that is where every competitor with a content budget and a subscription to the same tools has already piled in. The gap also compounded with time rather than staying fixed: Digital Applied measured it at 14% after three months and 31% after sixteen, meaning content that looked fine at launch kept losing ground as more competing material piled onto the same terms. The full methodology is worth reading before you apply this to your own roadmap, since the effect size will move with how contested your specific category already is.

The one piece of good news in that dataset is specific and actionable: AI-assisted content with real human editing nearly closed the gap against fully human-written work at every difficulty tier. The penalty sits on unedited AI output, not on AI involvement in the process. That reframes the fix. It is not "use less AI." It is publish less, edit harder, and stop sending unedited drafts at your hardest, most contested terms.

THE UNCOMFORTABLE PARTNone of this is a story about bad actors flooding the internet with spam. It is a story about competent teams, using the same tools, reaching the same conclusion at the same time, and collectively erasing the advantage any one of them expected to get from moving first. Saturation does not require anyone to do anything wrong. It only requires everyone to do the same right thing at once.

What Still Works Despite the Saturation

Trust concerns are rising in lockstep with volume, and that is exactly the terrain where sourced, structured work keeps an edge. An IAS and YouGov survey from October 2025 found 56% of respondents worried about a spam-heavy content experience and 52% worried specifically about unverified sources — both up from prior waves. Readers and AI engines are converging on the same instinct: skepticism toward content that cannot show its work. That instinct is exactly what a saturated field cannot fake at scale, because faking it well is slower and more expensive than the volume play it is supposed to replace.

1Fewer, sharper assets beat more, thinner onesDigital Applied's finding holds the fix inside it: cut volume at your highest-difficulty terms specifically, and put the editing budget you freed up into fewer pieces that actually clear the human-edited bar.
2Comparison and verdict formats keep their edgeContent that states a clear position and cites its claims is structurally harder to commoditize than narrative explainer content. We broke down why comparison and alternatives pages keep winning outsized attention, and the same structural advantage — a stated verdict, sourced claims, a real table — is what saturation erodes slowest.
3Named sourcing is now a ranking input, not just an ethics footnoteThe signals that earn both reader trust and AI citation overlap almost completely: named authorship, real credentials, and claims that trace back to something verifiable. That is expensive to fake and expensive to compete away at scale, which is exactly why it still works.

One more number worth sitting with: a Typeface survey found that 73% of teams using AI agents have already cut their agency spend on content production. Read that as proof that strategy is dead and you will misread your own budget. Raw production — the drafting, the formatting, the volume — is the layer that just got commoditized, and the market is pricing it accordingly. The layer that is not getting cut is architecture: deciding which few pieces are worth building at all, and for whom. That is the work behind hub-and-spoke content mapped to an actual buying committee instead of a keyword list, and it is the same discipline that carried Johann Wolff's four-year content partnership to a 186% organic revenue gain without trying to out-publish a saturated category.

Do this next: pull the list of everything your team published in the last two quarters and sort it by keyword difficulty. Anywhere you sent unedited AI drafts at medium or high-difficulty terms, you are sitting inside the 22-to-41% gap and competing against a saturated field that grew while you were publishing. Kill the thin duplicates, put a real editing pass on what survives, and redirect the freed-up budget into fewer pieces built like arguments — a stated position, sourced claims, a structure an engine or a skimming reader can lift without reading the whole page. Volume was never the fix. It was the thing that made the fix harder to find. The teams that come out ahead this year will not be the ones that published the most. They will be the ones that noticed the room got loud and quietly turned their own volume down to say something worth hearing.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.

JB
Josh BernsteinMANAGING PARTNER, SOMETHING INC.

Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.

Free consultation

Let us be the last SEO agency you ever work with

A 30 minute call and a free audit of your SEO and GEO position. You keep the findings either way.