AI is the new search. With ChatGPT showing product recommendations in the conversation, the focus is shifting from standard SEO to Shopify AI SEO.
Some stores already figured it out and have optimized their product descriptions for LLMs and Google’s AI overviews.
Here is how they did it, and how you can do it too.
We asked LLMs what they preferred
To understand what LLMs look for in a product, I went directly and asked them the following question.
“As an AI that often recommends users products to purchase, which Shopify product pages would you find easiest to read – and therefore most likely to recommend to shoppers? Find me 10 real examples across five industries of product pages that you would recommend to shoppers in a real-life situation.”
Chat GPT replied with the following.
Perplexity followed
It seems they both use the same criteria. A well-optimized product description that fits into an easy-to-scan structure is what they look for the most.
Traditional SEO description vs LLM Optimized product descriptions
All of search today is Semantic. In fact, Google is a gigantic semantic search machine.
This means that every search engine now aims to provide a comprehensive answer to user questions, rather than searching for exact or closely related keyword matches. The topic is the focus, not the keyword.
With this logic, your product pages need to clearly communicate to crawlers (and humans) what your product is, who it is for, and why people should buy it.
The table below effectively summarizes what LLMs are looking for in your product page.
| Standard Product Description | LLM-Optimized Product Description | |
|---|---|---|
| Focus Shift | Keyword-focused; lists features, sizes, and specs | Focuses on meaning, context, user intent, and outcomes. Helps AI and search engines understand what the product is for, who uses it, and why it matters |
| Structure | Short paragraphs or bullets, sometimes dense with keywords | Clear sections – title, intro, transformation bullets (benefits/outcomes), long description, Q&A, reviews, structured data hints |
| Language & Tone | Marketing jargon; adjectives like “premium” or “amazing” | Plain, human-friendly language, focused on practical outcomes |
| Semantic Context | Mentions keywords repeatedly (e.g., “kids guitar, children guitar”) | Uses semantic terms – synonyms, related topics, and phrases that reflect real search intent (e.g., “guitar for small hands, guitar for beginners”) |
| AI & SEO Friendly | Optimized for Google only via keyword placement | Optimized for both AI assistants (ChatGPT, Perplexity) and search engines. Includes context cues, Q&A, and structured info for better AI recommendations |
Shopify AI SEO examples from live Shopify stores
Below are examples from live Shopify stores you can learn from before rewriting your own. Notice how the same pattern repeats for all but is adapted to the industry, product, and audience.
Example 1 – Pourri
The selling point of Pourri lies in their mission, ‘eliminate funky odors without funky ingredients.’ Targeted towards pet owners, the company understood the concerns and pain points of their audience and answered all their questions in their PDPs.
What makes this product description stand out as LLM-friendly is how well-structured the page is and how it reinforces Semantic SEO with various FAQ style sections without repeating itself.
Example 2 – F-Musiikki
One of our clients, F-Musiikki, is a leading Finnish music instrument retailer that follows all Semantic SEO best practices with its Product descriptions.
F-Mussikki optimized PDP page
Their average order value (AOV) increased by 130%, and the conversion rate improved by 30% after moving from Magento to Shopify Plus, with the integration of Verbalic – our AI framework that automatically rewrites LLM-friendly PDPs.
You will notice how the product description includes a blend of specs, outcomes, and benefits, none of which are there randomly. In fact, asking ChatGPT for advice on a ‘guitar for little kids’ will show results in the LLMs suggesting a guitar based precisely on the specs and outcomes F-Musiikki mentions.
Example 3 – Neural DSP
Neural DSP is an example of a headless build where the front end of the PDP is built separately, and Shopify acts as the eCommerce engine.
The Nano Cortex product page is powerful for LLMs, as it answers multiple user queries about how, why, when, and what to use to achieve the desired guitar tone in various settings.
Example 4 – Beardbrand
All of the production descriptions for Bearbrand include related topics and phrases that reflect real search intent (Prevent beard itch, stop bear dandruff, etc.)
In the context of semantic SEO, the PDP covers the main concerns of buyers (safety, how to use, ingredients, and effect).
Example 5 – JB HI-FI
The Australian retailer JB HI-FI is another excellent example of a structure that thoroughly answers shoppers’ questions and satisfies LLMs.
Example 6 – Potgang
Potgang is a unique example of using the home page and PDP being one. The company sells only one product with different variations, so it technically needs just one landing page.
Why does this work for them?
By using the heading ‘GROW FRUIT + VEG FROM HOME’, they focus their content on answering the most common query that users have on growing fruits and vegetables, rather than the product itself. For this specific product, the specs are not as relevant as they are for a tech product, and the ‘how to’ content helps LLMs recognize the product as a solution to a problem.
Simply put, the landing page structure allows Potgang to give better, comprehensive answers to user questions, and LLMs love that.
The risk with this solution is that Google might not recognize your page as a product page based on its structured data. Applying all SEO best practices is needed to ‘compensate.’
Example 7 – Cotopaxi
Cotopaxi has a simple, focused product description supported by an image-heavy PDP.
The description clearly explains why this product is the right choice and for whom it is (specifies use cases for different laptop sizes). The rest of the copy on the page emphasizes the USP of the product – its eco-friendly and sustainable nature.
Before & After Examples of LLM Optimized Shopify product descriptions
For the final three examples, let’s transform a product description written for traditional SEO into an LLM-optimized description ourselves.
It’s not as simple as telling an AI what to do, even though we have a ready prompt for you. Prior to editing, some research is necessary.
Shopify semantic SEO keywords search
Semantic SEO research is different than normal keyword research. While normal keyword research focuses on a keyphrase and its variations, Semantic SEO includes related subtopics that are connected to the main topic.
Put simply, your product description needs to be comprehensive from all aspects and answer all questions a shopper might have.
The table recaps the methods covered and some other tools you can use. You can also look into the detailed procedure in our guide on rewriting Shopify product pages. In short, it looks like this
Keyword/semantic research → prompt → rewrite → structured data → publish → measure.
| Method | How to Use It | Notes |
|---|---|---|
| SERP scan | Look at the top 10 Google search results for “Guitar for little kids.” Copy bolded or highlighted phrases. | Captures Google’s semantic understanding of the topic. |
| People Also Ask (PAA) | Open PAA box on Google results. Collect 5 10 common questions. | Great for Q&A blocks in PDPs. |
| Google Suggest | Type the query + letters (e.g., “guitar for little kids a…”) | Reveals related search intents. |
| Keyword / AI tools | Tools like KeywordTool.io, Surfer, MarketMuse, or AnswerThePublic | Generates semantically related terms and subtopics. |
| Competitor analysis | Check top-ranking product pages. Note phrases, feature language, and buyer-centric terms. | Helps fill gaps your content might miss. |
The semantic AI prompt for LLM friendly product descriptions
All the examples below were rewritten using the following prompt, considering the research on the products/industries/audience, as well as AI SEO tips and tricks from our previous article.
Rewrite my Shopify product description using semantic SEO principles so that it’s optimized for ChatGPT, Perplexity, all LLMs, and also Google indexing. Focus on clarity, context, and structured information.
Here’s my current description:
Include these elements in the rewrite:
- A clear title under 70 characters that mentions (Add here the findings from your research)
- A short intro paragraph that explains:
- What the product is
- Who it’s for
- When it’s used
- Three “transformation” bullets that describe outcomes (not just features).
- A short Q&A section (3 real customer-style questions and concise answers).
- Write naturally, in plain English. Avoid buzzwords like “premium” or “amazing.”
My product details:
(based on your findings and knowledge of the product/industry/audience)
- Product name:
- Type:
- Key materials/features:
- Audience:
- Use case:
The tone should be clear, friendly, and informative, suitable for suitable for eCommerce shoppers and AI discovery tools.
How we got to this prompt is just as interesting and helpful as the prompt itself.
Example 8 – All Stars
The first is an example of a straightforward product page structure we ran through our prompt. The only edits we made were to insert a few instructions about the audience (young people, creative professionals) and usage. (All-day wear, daily wear)
While the first description includes specs and outcomes (benefits), the improved variation is more scannable for both machines and humans, clearly outlining its target audience.
From a semantic perspective, the improvement involves changing the general wording “Sneaker, comfort, premium” to more specific terms like “all-day wear,” “heritage,” “everyday comfort,” and “creative professionals.”
Example 9 – Tummy Focus
The second example is from a health supplement. In this scenario, we edited the prompt to include the specific ingredients of the supplement and fact-checked the outcomes, considering the sensitive nature of the product.
Note: these are not real benefits or ingredients; they’re for demonstration purposes only.
Example 10 – Scissor Jack
The example below features a spec-heavy item that already has a traditional SEO optimized PDP.
The focus of the description has now shifted from pushing the main keyword ‘Off-road Scissor Jack’ and listing specs to expanding on related terms and the benefits. The table below offers a detailed breakdown.
| SEO Element | Standard Version | LLM-Optimized Version | Key Change | Specific Example |
|---|---|---|---|---|
| Keywords | Exact-match terms | Semantic + contextual terms | Expanded from keywords to meaning and intent | “3.5-ton scissor jack” → “heavy-duty 3.5-ton off-road jack for Sprinter vans and other platforms” |
| Structure | Basic headings | Intent-based sections & cues | Clear sections for AI readability | Overview, Specification, Instructions → Overview, Key Features, How It Helps, Q&A |
| Specs | Listed plainly | Linked to benefits & scenarios | Features tied to use cases | “Extended height 17.5” → “Extended 17.5” height lifts your vehicle safely, ideal for uneven terrain” |
| Tone | Informative | Conversational + outcomes | More human-friendly and scenario-focused | “Includes crank handle, ratchet, and carry bag” → “Carry bag and crank handle included so you can safely lift and transport your vehicle anywhere” |
| Search Optimization | Google-focused | AI + Search + Voice optimized | Optimized for LLMs and voice search too | Basic specs → Added context for “off-road lifting”, “Sprinter compatibility”, “safe vehicle maintenance” |
| Compatibility Info | Basic mention | Specific platforms & contexts | Detailed compatibility and usage context | “Compatible with Sprinter and other platforms” → “Works with Sprinter, camper vans, and medium-duty trucks requiring 3.5-ton lift capacity” |
| Schema Signals | Absent | Implied through formatting | Structured info signals for AI | Ingredients / materials & weight now in structured bullet points |
| Use Cases | Not addressed | Highlighted throughout | Connects features to real-world scenarios | “Billet aluminum dual-sided adapter” → “Dual-sided adapter fits both axle and flat surfaces for safer and versatile lifting” |
Recap – The key lies in automation
The examples above, except for a few, consist of websites with hundreds or thousands of SKUs.
While it can take a team to export manually, rewrite (with or without AI), and copy again in the back end, we have built our own system, Verbalic, that automates the process for you and continually updates the content.
Book a call with our team to find the best way to have your products featured in Google AI Overview and all LLMs.
