Updated
Human writing is making a comeback, and I would start with product descriptions.
Since generative text tools went mainstream, the content world has been flooded with machine-written copy, blog posts, product descriptions, social captions, email newsletters, and if it can be written, someone has tried to automate it. The result is a sea of content that is competent on the surface and empty underneath. Nowhere is that more visible than on a product page, where ten stores selling the same SKU now carry ten near-identical paragraphs that came out of the same model with the same prompt.
I use AI in my workflow most days, and what I want on a product page is the thing the tools cannot replicate, a human being who has held the product and has something to say about it.
The Glut of Product Copy That Sounds the Same
When the same writing tools sit on desks across the industry, the output starts to sound the same. Ask Claude or ChatGPT to describe a product or explain a business topic and you get clean, structured, hedged prose that says nothing original. It is the content equivalent of elevator music, present, inoffensive, and forgettable.
Google has noticed, and readers have noticed. The glut has made original human writing more valuable, and when the bulk of what a shopper reads sounds the same, the description that sounds like a person with an opinion stands out, which is a ranking argument as much as a taste argument.
What Human Writers Bring to a Product Page
A human writer brings lived experience, and I can tell you what it was like to launch the first Magento B2B site because I was there. I can describe the feeling of watching a deploy fail at 2 AM because it has happened to me. AI can write about those topics, but it has not lived through one of them, and the writer who has unboxed the thing, worn it on a long run, or returned it because the zipper caught knows what to say in the first line.
Human writers also bring judgment, which is knowing what to leave out as well as what to include. AI fills whatever space it is given, so a person has to decide what the space is for, and that call separates a description that is comprehensive from a description that is useful.
And human writers bring voice, the imperfect, opinionated kind that makes a shopper feel like someone on the team is talking to them across the counter, which is a different thing from the “brand voice” paragraph in a marketing brief.
How to Write Product Descriptions for SEO With a Human in the Loop
Start with the phrase a shopper types, and write the opening line the way a person who has used the product would answer it. The search intent behind “how to write product descriptions for SEO” tends to be a store owner or category manager staring at a few hundred SKUs with thin or copied copy, so the method has to scale without turning into the elevator music above. The split below is the one I use.
- The human writes the first paragraph and the objection. The opening sentence names the product and the query the way a shopper phrases it, and the second or third sentence answers the one doubt that stops the purchase, the fit, the weight, the noise, the thing the spec sheet will not admit. Only a person who has handled the product can write that line with any authority, and Google’s quality guidance asks for that line.
- AI drafts the structured block. The dimensions, materials, compatibility, care, and variants live in your PIM or your spreadsheet, and a model is good at turning that data into clean, consistent copy across hundreds of pages. Pair it with Product structured data so the machines reading the page get the same facts, which is the same argument I made about product data for AI channels.
- The keyword goes where a reader would put it. Title, first line, one subheading if the page has one, image alt text, and then stop. The earlier piece on writing SEO-friendly product descriptions still holds on placement, though the penalty for sounding like the other nine stores has grown since it ran.
- Read it out loud before it ships. If it sounds like a spec sheet with adjectives, the first paragraph needs a rewrite. Our guide to product descriptions that convert covers the persuasion side, the half of the page a person writes.
What Google Says About Machine-Written Copy
Will AI copy rank? Google’s own guidance settles that question better than any agency opinion. Its guidance on AI-generated content says its ranking systems “aim to reward original, high-quality content that demonstrates qualities of what we call E-E-A-T” (expertise, experience, authoritativeness, and trustworthiness), and that using AI to generate content “primarily to manipulate search rankings” is against its spam policies. Its helpful-content self-assessment asks whether a page demonstrates first-hand expertise, and gives “expertise that comes from having actually used a product or service” as the example.
Read those two lines against a page of model-generated descriptions for products nobody on the team has touched. Google allows the tools and requires the experience, and the experience has to come from someone who has used the product. I wrote up the operating version of that stance in our AI writing policy.
The Right Balance
I am not advocating for throwing away AI tools. Throwing them away would be like arguing against spell check because it is not “authentic.” AI is a strong drafting and research tool, and it accelerates the mechanical parts of writing, the variant copy, the spec tables, the first pass at a category page.
The thinking, the perspective, the editorial choices, and the voice need to come from a person. The content that performs best is content where AI handled the scaffolding and a human made it something a person wants to read. Content Cucumber kept its writers while the rest of the industry was letting theirs go for that reason, and the decision pays off fastest on product description work.
Write Your Story
Business owners, category managers, and product experts with experience to share have writing that carries more weight than it did before the generative tools went mainstream. The bar for generic content has been obliterated by AI, and the bar for experience-driven content has moved up along with the reward for clearing it.
Write about what you know. Write about what you have seen. Write about the product that failed in the field and what you changed. Content built from your own experience is content AI cannot produce, and readers and search engines are hungry for it.
Humans of the world, write, and let the machines handle the rest.
When the catalog is too big for the team to write by hand, Content Cucumber’s product description writing service pairs human writers with AI tooling so the copy sounds like a person wrote it, because a person did.
Explore our writing services
Content Services →Keep Reading
Frequently Asked Questions
Google's guidance on AI-generated content says its ranking systems aim to reward original, high-quality content that demonstrates expertise, experience, authoritativeness, and trustworthiness, however it is produced, and that generating content primarily to manipulate search rankings violates its spam policies. AI-drafted descriptions can rank when a person with first-hand experience of the product writes the opening line and answers the buyer's objection.
Put the search phrase in the title, first line, one subheading, and image alt text the way a shopper types it, have a person who has used the product write the first paragraph and answer the one doubt that stops the purchase, and let AI produce the structured spec block from your product data.
The structured block, meaning dimensions, materials, compatibility, care instructions, and variant copy drawn from the PIM or spreadsheet, paired with Product structured data. The opening line, the objection, and the voice come from a human who has handled the product.
In the title, the first sentence, one subheading if the page has one, and the image alt text, then stop. Repeating it beyond that reads as stuffing to shoppers and to Google.
Google's helpful-content self-assessment asks whether a page demonstrates first-hand expertise, and gives expertise that comes from having used a product or service as its example. Shoppers can also tell the difference between copy written from a spec sheet and copy written by someone who used the item.
