When Everyone Can Publish 500 Articles, What Is Content Worth?

There is something slightly strange happening in digital marketing right now.

For years, companies were told they needed more content. More useful articles. More landing pages. More answers to more search queries. More opportunities to appear in Google.

The problem was always the same: producing good content took time. Someone had to know the subject, research it, write it, edit it, and, ideally, someone who actually understood the business had to make sure the finished piece was more than a collection of facts copied from somewhere else on the internet.

Then generative AI changed the economics completely. Today, an article that might once have taken a writer and an editor several days can be produced in minutes. A company can generate ten ideas before lunch, fifty articles by the end of the week, and, with the right automation, potentially thousands of pages without significantly increasing its content budget.

It is easy to see why this is exciting. It is also easy to see why so many companies are about to make the same mistake.

They have discovered that AI makes it incredibly cheap to produce content. They have not necessarily stopped to ask whether the world actually needs another piece of it. And those are two very different questions.

Content Became a Commodity

Content Became a Commodity

The internet has always had an interesting relationship with quantity. When search engines were young, there simply was not enough information available online. Even basic content could be valuable because it was hard to find. SEO grew out of that gap: businesses learned how to structure pages, target queries and make their information easier to discover.

But the economics were different then. Publishing twenty genuinely useful articles was an investment. Publishing two hundred was a much larger one. That natural limit gave content a certain scarcity.

Generative AI has removed much of that limit, and when the cost of producing something approaches zero, the thing itself tends to become less scarce, and therefore less valuable. We are already seeing the beginning of this with written content: thousands of companies now publish articles that are technically competent, grammatically clean and perfectly structured, but ultimately tell the reader very little they could not have learned somewhere else.

The sentences are different. The information often is not.

Traffic Isn't Value

Traffic Isn't Value

The temptation is understandable. If ten articles a month have been slowly growing your organic traffic, and AI now lets you publish one hundred, why wouldn't you? If one article has a chance of ranking, one hundred give you one hundred chances. In some cases, they genuinely do: a new page can rank, a low-competition query can bring traffic, a large site can suddenly appear for hundreds of long-tail searches, and for a few months the graphs can look spectacular.

That is exactly what makes the story more complicated, because mass-produced AI content can produce real results, at least by one definition of results. The important question is not whether the strategy can produce traffic. It is whether it can produce lasting value.

This is something we explored in our article on SEO, AEO and GEO in 2026: a good SEO report should never stop at rankings, impressions or the number of pages published. Those metrics describe what happened in search. They do not necessarily describe what happened to the business.

Imagine a company publishes 300 articles. Organic impressions rise 70%. The site now ranks for hundreds of new queries, and the report shows beautiful upward graphs. But qualified leads increase by 3%, revenue does not move, and most new visitors arrive on informational pages, read nothing else, and leave. The company has succeeded at producing more search visibility. It has not necessarily succeeded at building a better business.

When every additional article costs almost nothing, it becomes extremely easy to mistake volume of production for progress.

Better Reasons to Rank

Better Reasons to Rank

There is a misconception that the future of search involves Google trying to detect whether a paragraph was written by ChatGPT. We don't think that is the interesting question, and Google's own guidance doesn't frame the problem that way either. Google does not prohibit the use of generative AI to create content; its guidance acknowledges that AI can be useful for research, organization and other parts of the process.

The problem begins when large numbers of pages are created primarily to manipulate rankings rather than provide something useful to people. Google describes this as scaled content abuse, and its guidance explicitly includes cases where generative AI is used to produce many pages without adding value.

That distinction matters, because the real question was never "was AI involved?" It is "why does this page exist?" If the answer is that someone identified a search query and wanted another URL to rank for it, that is a weak foundation. If the answer is that the company had something genuinely useful to explain, investigate or document, AI can be an excellent way to bring that information into the world.

Knowledge vs. Words

Knowledge vs. Words

Consider two companies. The first has spent fifteen years working in its industry, completed hundreds of projects, and learned what works and what does not. It has internal processes, customer stories, research and opinions that developed through actual experience. Give that company AI, and the results can be extraordinary: it can turn a long expert interview into a detailed article, analyze customer feedback for patterns, document the decisions behind a complicated project, and help an editor find gaps that were easy to miss. AI multiplies that company's knowledge.

Now consider a company with very little original knowledge, no research, no case studies and no proprietary data. It simply wants more pages, so it asks an AI assistant to write an article about a topic. Then another. Then another. Soon it has 500 pages. That company has not multiplied its knowledge. It has multiplied its words.

That is the distinction we expect to hear far more about.

Why Experience Matters More

Why Experience Matters More

AI can generate text extraordinarily quickly, but people do not generate new knowledge at the same speed. A company can ask AI to write 100 articles about website design tomorrow; that does not mean there will be 100 new insights about website design tomorrow. Most of those articles will inevitably be built from information that already exists: one explains it, another summarizes it, another rewrites it, another combines five existing pieces, another adds an FAQ, another simply changes the tone. Eventually, the internet fills with different documents containing essentially the same ideas. The amount of text grows exponentially. The amount of original information does not, and that creates an unusual situation: information becomes abundant while originality becomes scarce.

This is one reason Google's concept of E-E-A-T includes Experience. Experience is difficult to manufacture because it comes from something that actually happened. A company can say it understands website redesign, or it can show what happened when it redesigned one. Those are not equivalent. A generic article explains the principles of a successful redesign; a real case study can show the original problem, the constraints, the rejected ideas and what happened afterward. The second contains something the first cannot easily reproduce: evidence of experience, which is exactly what we discussed in why content with a real point of view sells better than polished AI copy.

This is also why we expect the strongest content to become more personal, not more automated, at the higher end of the market. It will increasingly begin with something AI did not create: an interview with a specialist, a real project, an experiment, a customer conversation, a dataset, a mistake, a lesson learned, a strong professional opinion. AI can then process all of that material: organize a three-hour interview, compare data, identify themes, suggest structure and turn one piece of research into several useful formats. That is a very different use of AI than asking it to invent another 2,000 words about a subject that has already been covered a thousand times.

The Liability of 500 Articles

The Liability of 500 Articles

There is another problem with mass publishing that has very little to do with Google: eventually, someone has to manage all those pages. Which ones are your best pages? Which are outdated, which overlap, which answer essentially the same question? Which should be merged, rewritten or simply removed? At ten articles, these questions are manageable. At five hundred, they become a content-management problem of their own. The site gets larger, but not necessarily stronger, and in some cases it becomes harder to understand.

A website is not automatically more authoritative because it contains more URLs. Google itself explicitly warns against the assumption that producing a large quantity of pages makes a site more useful or relevant, and this has a human dimension too. Most readers cannot tell you exactly why a paragraph feels AI-generated, but they can often feel when something is generic: the introduction sounds familiar, the examples are interchangeable, everything is technically correct but nothing is memorable.

That matters enormously for businesses selling expertise. If you sell a commodity product, generic content may not hurt much. If you sell architecture, consulting, software development, strategy, design or any other expertise-driven service, your content is also part of your reputation. A thousand generic articles can make a company look less distinctive, not more.

AI Search Rewards Originality

AI Search Rewards Originality

Search itself is changing. Google's search experience increasingly includes generative features, and it has begun providing Search Console reporting for visibility in AI Overviews and AI Mode. The user is no longer always presented with ten blue links and asked to pick one; increasingly, the system synthesizes information from multiple sources and constructs an answer. We covered what that shift means in practice in why your business can become invisible to AI search, and how to fix it.

That creates a different kind of competition. It is no longer enough to ask "can we rank this page?" We also have to ask "does this page contain something worth using as part of an answer?" A generic article that repeats information found everywhere else has little to distinguish it. A page with original research, first-hand experience, unique data or a genuinely useful framework gives the system something far more interesting to work with. This is one reason we don't see AEO and GEO as a replacement for SEO. They are an extension of the same idea: make valuable information easy to discover, understand and trust.

Factory vs. Knowledge System

Factory vs. Knowledge System

There is an economic paradox here. When content was expensive, companies had an incentive to make each piece count. Now that it is cheap, they have an incentive to produce more of it. But when everyone can produce more, quantity loses much of its competitive advantage, and the advantage moves upstream, toward the source of the information: who has the data, the experience, the customer, the project, the mistake, the lesson, the result to demonstrate? That is where the real scarcity is, and it is why we expect the value of original expertise to rise as the cost of producing generic content keeps falling.

There is a subtle but important difference between a content factory and a knowledge system. A content factory asks how many pages it can produce this month. A knowledge system asks what the company knows that would be useful to other people. The first starts with keywords and measures production. The second starts with expertise and measures usefulness. Both can use exactly the same AI tools, which is why we don't think this is a debate about AI content versus human content, as we touched on in why most businesses are using AI in marketing the wrong way. The real question is whether AI is being used to create more information, or to make valuable information more accessible.

Using AI the Right Way

Using AI the Right Way

So does mass AI content actually work? Sometimes. There are still situations where a large number of pages generates search visibility: some queries are easy, some niches have weak competition, some sites have enough authority that even mediocre pages get traffic, and algorithms are never perfect. A company can publish hundreds of AI-generated pages and watch traffic increase. The problem is what happens next. Does the traffic remain? Does the content keep performing? Does the site build authority, does the brand become more trusted, does the traffic convert? A tactic can work without being a good long-term strategy, and we would be especially careful with any approach whose implicit assumption is "Google hasn't figured this out yet." That is not a durable advantage. Google's current guidance already makes clear that creating large numbers of pages primarily to manipulate visibility is not a sustainable direction, so the better question isn't whether a loophole exists today, but whether the strategy still makes sense once it closes.

Our answer: use AI, absolutely, but at the right point in the process. Instead of Keyword → Prompt → Article → Publish, we would much rather see Experience → Research → Data → AI → Expert review → Editorial refinement → Evidence → Publish. The difference is subtle but fundamental. In the first model, AI is the source of the content. In the second, it is the production system around human knowledge, and that is where it becomes genuinely powerful.

One Project Beats Twenty Articles

One Project Beats Twenty Articles

Suppose an agency completes an interesting redesign. There is usually enough material inside that one project for an entire content cluster: the original problem becomes an article, the design decisions become another, the technical challenges become a third, the analytics become a case study, the lessons become a guide, the methodology becomes a service page, and the project itself becomes part of the portfolio. One real piece of work can produce an entire body of original information, which is a far more interesting way to think about content. You are not manufacturing articles. You are documenting the work your company is already doing.

This is where AI becomes an unfair advantage for companies that already have expertise. An architect can record a conversation about a project and turn it into an edited case study. A software company can analyze years of support tickets and surface recurring problems. An agency can turn its project archive into a library of real lessons. A consultant can transform hundreds of pages of research into a structured knowledge base. The limiting factor is no longer the ability to write. It is the quality of the underlying material, and that changes the role of the content team: the valuable skill is no longer "can you write 2,000 words?" It becomes "can you find something worth saying?"

The Test Before You Publish

The Test Before You Publish

There is one test we find particularly useful before publishing any AI-assisted article: would this page still be worth creating if Google did not exist? Not because Google is irrelevant, it obviously isn't, but because the question forces us to think about purpose. If the answer is "yes, our customers need this information," good. If it's "yes, we have unique experience to share," even better. If it's "yes, this contains research or data we collected," excellent. But if the honest answer is "we found a keyword with search volume," the page probably needs a better reason to exist.

The fundamentals haven't disappeared. Technical SEO still matters, search intent still matters, internal linking still matters, structured data still has its place. But the environment around all of it is changing: Google has to distinguish between millions of pages, AI search systems have to decide which sources are worth incorporating into an answer, and users have more information than ever and less patience for generic content. Under those conditions, originality becomes a competitive advantage again, not as a writing trick, but as evidence that someone actually knows something.

The real scarcity is no longer content. It is knowledge.

AI has made the production of language incredibly cheap. That is a remarkable achievement, but it also means another perfectly competent article about "10 website design trends for 2026" is not particularly hard to produce. What is difficult is having something genuinely interesting to say because you have spent fifteen years doing the work. A real case, a real result, a real dataset, a real failure, an opinion that survived contact with reality. Those things cannot be generated by changing the prompt. They have to come from somewhere, and increasingly, that somewhere is the competitive advantage.

We don't think the future is a world without articles, quite the opposite: there will probably be more useful content than ever. But there will also be an enormous amount of material that exists mainly because it was cheap to produce, and the difference between the two will matter more with every year. Companies that treat AI as a replacement for expertise may find they can produce content faster than ever without becoming any more authoritative. Companies that treat AI as an amplifier of expertise can do something far more powerful: take everything they know and make it easier to find, understand and trust.

And that brings us back to the same question we should have been asking about SEO all along: are we creating pages because we want to rank, or because we have something genuinely worth finding? In the short term, those two things can look remarkably similar. In the long term, they are not, and we suspect the difference is about to become much harder to hide.