AI Content at Scale: Why Good Enough No Longer Works

Share

All Articles

AI-generated product content is now standard for large catalogues. The tooling works. The quality variance is the problem nobody planned for, and on a 50,000-SKU catalogue, even a 1% error rate is 500 wrong product descriptions.

Topics
Sarah ChenSenior Editor

—

The vendor case studies rarely cover this part.

AI content generation for product catalogues works. The speed advantages are real. A retailer that used to spend three weeks writing, reviewing, and approving product descriptions for a new range can now get a first draft in hours and spend the remaining time on quality control rather than generation. For large catalogues, that has been genuinely transformational for time-to-market and for teams that were drowning in backlog.

And then there is the other part.

Earlier AI tools had meaningfully higher error rates for product specifications: wrong materials, invented certifications, dimensions that did not match reality. Current frontier models, in 2026, have brought that down considerably, with informed estimates around 1 to 2 percent. That sounds like an enormous improvement—and it is. But on a catalogue of 50,000 SKUs, 1 percent is still 500 product descriptions with something factually incorrect. Not obviously wrong (not “this television can fly” wrong), but quietly wrong in the specific detail that makes a product description useful. The wrong material composition. A certification that does not exist. A size the product does not come in. “Italian leather” on something that is not Italian and is not leather.

The model is optimising for “plausible product description”, not “accurate product description.” Those are not the same thing, and the distinction matters when customers base purchasing decisions on what you have written.

The Blandification Effect

The specification problem gets you into trouble with trading standards and customer returns. But a subtler quality problem is arguably more damaging for brand-led retailers: AI copy regresses to the mean.

The training data for these models is the entire corpus of product description writing on the internet, most of which is neither particularly good nor particularly distinctive. The output reflects that. Quirky brands produce content that sounds like every other quirky brand. Technical brands lose the specific precision that makes their content credible to their actual audience. The distinctive voice someone spent years developing gets sanded smooth because the model has learned that “smooth” is what product descriptions sound like.

This is not a problem you fix by prompting alone, though prompt engineering helps at the margins. The deeper issue is that brand voice is accumulated institutional knowledge: knowing which words the brand would never use, knowing the rhythm of a sentence that fits the aesthetic, knowing when to be funny and when not to be. That knowledge does not transfer cleanly to a context window.

Some brands have addressed this by building brand voice frameworks into their generation workflow: style guides, approved vocabulary lists, examples of copy the brand would never write, and specific guidance on handling technical specifications. This helps significantly with voice consistency. The brands that make it work have usually spent years producing good human-written copy and have enough of it to establish real patterns. The ones starting from scratch with no archive and no voice documentation are generating into a void.

The Editorial Gap

The thing that often gets cut when AI content generation goes into production at scale is the editorial layer. The logic is understandable: if you generate content ten times faster, you should be able to review it ten times faster too, so net headcount stays the same. That is not how it works in practice.

Reviewing AI-generated content for factual accuracy requires domain expertise, not just editorial instinct. A human reviewer who is not a subject matter expert can catch obvious errors and clunky phrasing, but will not necessarily catch the confidently stated wrong specification. You need someone who actually knows what the product is meant to do and what it is made of: a buyer, a product manager, or someone who has been briefed specifically. Not just someone with good grammar and a deadline.

The retailers who are getting this right have generally built their workflows around an exception-based model. AI generates; a rule-based system flags anything that deviates from expected ranges (materials that do not match the product type, dimensions that seem implausible, certifications that require verification); and human reviewers focus their time on flagged items rather than reading everything. The unflagged items still get spot-checked, since the error rate is low enough that random sampling catches most issues, but the systematic review is targeted.

This workflow is more expensive than “AI generates, light human review, publish.” It is also considerably cheaper than the original all-human process, and it produces higher-quality output than “AI generates, auto-publish”—which some teams are running in the name of speed.

The Search and AI Visibility Dimension

There is a machine-readability angle to this that I wrote about last week. AI-generated content with errors and inconsistencies scores poorly on the machine readability metrics that determine whether AI discovery systems surface your products accurately. This creates a compounding problem: not only do you have wrong product descriptions visible to customers who land on your site, but those same descriptions degrade your visibility in the AI-mediated discovery layer driving an increasing share of your traffic.

A product page that says “Italian leather” when the product is synthetic microfibre gets served to people looking for leather goods, who then return it, or worse, review it poorly. It also gets served to people asking an AI assistant for Italian leather products. Which is a separate version of the same problem.

The quality problem radiates outward from the content itself.

What This Actually Requires

None of the solutions here are exotic. They are mostly just discipline:

Clean source data in the PIM before generation. If your raw product data has inconsistencies and gaps, the AI will fill those gaps with plausible-sounding invented content. The generation step cannot fix upstream data quality problems.

Verification workflows proportionate to product category risk. High-value products, technically complex products, and products with regulatory requirements (food, cosmetics, children’s products) need more rigorous review. Sports accessories need less.

Periodic audit of published AI-generated content against actual product specifications. Not just before publishing but after, because product specifications change and generated content does not update itself.

Brand voice documentation specific enough to be useful in a prompt. Not “write in a warm, conversational tone” but examples of approved sentences, lists of words the brand would and would not use, and specific guidance on how the brand handles technical specifications.

The AI content machine is running. For most retailers at scale, it needs to keep running; going back to fully manual content production is not the answer. But “it is running” and “it is running well” are meaningfully different, and the gap between them is not closing on its own.

Tags

Stay Connected

Follow LLCommerce on LinkedIn

Get the latest AI commerce insights, analysis, and industry news delivered to your feed.

Large Language Commerce

About the Author

Sarah Chen
Sarah Chen

Senior Editor

Sarah covers the intersection of AI and retail, with over a decade of experience in technology journalism. Based in Bangkok, Thailand — and will explain at length why that’s actually the best place to cover e-commerce if you’ll let her.

Read more