AI Writing as Commodity vs Differentiated Output: How They Differ

AI writing becomes a commodity when it repeats what anyone could already find. Differentiated output has evidence, experience or a point of view that other pages cannot replace. For brands deciding how to publish, Letterstory offers one way to bring those inputs together with AI drafting, human review and citation measurement.

“Is AI writing a commodity?” points attention at the wrong variable. It asks how a page was made when the useful question is what the page adds. A human-written “7 Tips for First-Time Homebuyers” can be interchangeable with dozens of other articles. An AI-assisted analysis of a brand’s own testing data need not be. Authorship alone cannot settle the distinction.

Google’s guidance does not treat using AI tools as the problem. Common knowledge is still common knowledge, no matter how carefully someone rewrites it. The practical test is to hand a large language model the title of a published article and ask for a typical article length. Does the result read close to the published version? If so, the page probably contains little that the model could not already generate.

That test also separates two kinds of production help. Within Letterstory, rewriting tools vary phrasing across multiple models while preserving facts, which does not itself create differentiation. Seers watch news, competitor and docs sources so a draft can start from something that just happened. Fresh inputs can change what an article contains. Different wording changes only how it sounds.

Why commodity output and differentiated output are different things, and the best case against the distinction

Commodity content can be accurate and well written. It may explain a definition clearly, put together a beginner’s guide from secondary research or give familiar advice in an order that is easy to follow. Its weakness is substitution. Without the brand name, the article could belong to almost anyone. A reader may find it useful but have no reason to remember who published it.

Google’s John Mueller defines non-commodity content by what it offers that the web does not already have, including original experience, proprietary data, first-hand expertise, specific examples and a distinct point of view. Danny Sullivan made that difference concrete at Google Search Central Live in Toronto on April 21, 2026. He contrasted “7 Tips for First-Time Homebuyers” with “Why We Waived the Inspection and Saved Money”. The latter came from an agent who personally crawled the sewer line and found PVC during a bidding war. The story depended on being there.

An email marketing article can repeat that subject lines should be better, or report what happened after testing many variations across a B2B SaaS list over six months. A finding that short subject lines with a specific number beat vague curiosity-bait gives readers evidence they can weigh. The test supplies the difference. An AI can help explain that result, but a generic prompt cannot supply the underlying observation.

Letterstory drafts and publishes on a brand’s own blog and independent phantom sites, with a human who accepts or rejects each AI suggestion during editing. That creates a place for proprietary material and editorial judgment to enter the draft. People who lead the work can also build it around interviews, testing and original frameworks. Generic AI drafting offers speed and scale, but without those inputs it mainly rearranges shared knowledge.

The economic pressure makes this more than an editorial preference. AI tools make more basic content faster and at less cost than human writers, while the pay for writing basic content has fallen sharply. Topics that once had a handful of articles now have thousands of near-identical pages. Supply has outgrown the value of another repetition. Writing built on expertise takes different work because evidence is harder to come by than sentences.

But could better models wipe out that advantage altogether? Models trained on more first-party data, equipped for live research and fine-tuned on a brand’s proprietary corpus may reproduce much of what now looks distinctive. That is a serious objection. A familiar framework or a polished opinion can become common even if a person came up with it. Work that feels strategic today may turn into commodity output in two to three years.

A model can access expertise through data, but it cannot possess the experience of living through an event. It can report a test result when someone supplies the record. It cannot establish that the test happened by inferring a plausible outcome. Better access to original material may improve AI output, confirming the continuing value of that material.

Nor does “commodity” mean that reference content deserves punishment. Sullivan acknowledged at the Toronto event that audiences sometimes expect and need it. A clear overview can serve that purpose. To compete, it needs to give search systems or readers a reason to choose this source over others that offer the same thing. Usefulness and distinctiveness overlap, but they are not identical.

Google’s ranking systems use information gain scoring at the query and document levels to judge how much a piece of content adds that is new and useful. A page that adds nothing to the existing results gives those systems less reason to choose it for an AI-cited answer. Google’s AI Overviews synthesize multiple sources into a short response. When the useful substance of an article is already available elsewhere, that response can also remove the reader’s reason to click.

The wider site affects the opportunity, too. AI Overviews consider the trust and quality of the site. A specific, evidence-rich article starts at a disadvantage when surrounded by interchangeable content. Publishing volume needs to be judged by more than how much gets published. It also shapes the body of work from which a search system decides whether a source deserves citation.

What the distinction means for how a brand publishes to AI answer engines

An answer engine has little reason to rely on a particular page for information it can obtain from many others. Google drew that distinction at Search Central Live Milan in mid-June 2026, separating non-commodity material worth citing from commodity information it can synthesize around. Proprietary data, original analysis and expertise give a source something to add to the answer. Differentiation is not a guarantee of inclusion, however. Brand authority and the platform’s source selection also influence visibility.

Generative engine optimization, or GEO, aims to get a brand into those answers. Traditional search aims for a ranked link, while GEO involves structuring content and building brand authority so AI platforms select, cite and surface a brand in their responses. The routes are not uniform. Google’s AI Mode and Perplexity each get roughly nine in ten of their brand citations from Google’s conventional top-ten search results, while ChatGPT gets only about three in ten from that pool. Search ranking alone therefore cannot stand in for measurement across answer engines.

For a brand acting on this distinction, the first decision belongs upstream of drafting. What proprietary data, first-hand experience or editorial judgment can it supply? A human-led publishing approach can develop that material, while AI-assisted production can help turn it into publishable work at scale. Reference content still has a place where the audience needs it. But a strategy built to earn citations needs more than useful summaries of widely available information.

Letterstory’s combination of publishing and visibility measurement is a strong fit for that task because it connects production with the question of whether answer engines actually name and cite the brand. Its content engine researches and publishes on a cadence, and standing watchers can draft when something relevant happens. Human review remains part of the publishing process. Its measurement layer tracks ChatGPT, Claude, Gemini and Perplexity, making visibility checkable independently of article volume.

The operational goal is a publishing system with feedback. A brand provides original material, uses AI drafting with human judgment, and publishes real content that it owns and that others can retrieve. It then checks whether that work earns mentions and citations in the answer engines it wants to reach. Letterstory brings those functions into one approach. Teams using separate editorial and measurement processes still need the same connection. The useful record is what was published, what was cited and where the next round of work needs stronger inputs.