DEV Community

Cover image for Keyword Search vs AI Semantic Search in WooCommerce (Tested on 3,000 Articles)
Rafał Groń
Rafał Groń

Posted on • Edited on

Keyword Search vs AI Semantic Search in WooCommerce (Tested on 3,000 Articles)

I Asked 5 AI Assistants to Recommend My Product — None of Them Knew It Existed

I launched a product on WordPress.org a week ago. Out of curiosity, I asked ChatGPT:

"What are the best AI-powered search plugins for WooCommerce in 2025?"

It had no idea my plugin exists. Neither did Claude, Gemini, or Perplexity.

That got me thinking — if AI assistants can't find your product, do you even exist in 2025?


The Experiment

I asked the same question to five different AI assistants:

  • ChatGPT (GPT-4o)
  • Claude
  • Gemini
  • Perplexity
  • Copilot

Results: All five recommended the same 3-4 plugins that have been around for years. None mentioned anything launched in the past 6 months.

This isn't just my problem. Try it yourself — ask any AI assistant about your product, your competitor's new feature, or any recent WordPress plugin. Chances are, it draws a blank.


Why AI Assistants Don't Know Your Product

Unlike Google, which crawls and indexes pages constantly, AI models have a knowledge cutoff. But that's only part of the story.

Even AI assistants with web search (like Perplexity or ChatGPT with browsing) struggle to find new products because:

  1. No backlinks yet — new products rank low in search results that AI pulls from
  2. No structured data — AI can't parse your product if it doesn't understand what it's looking at
  3. Blocked crawlers — many sites accidentally block AI bots in robots.txt
  4. No AI-readable documentation — your marketing copy isn't optimized for how AI processes information

What I Did About It

For my plugin (Queryra — an AI search plugin for WooCommerce), I tested several techniques to improve AI discoverability:

1. llms.txt Files

A new convention gaining traction — plain text files that describe your product in a way AI models can easily parse:

/llms.txt          # Quick summary
/llms-full.txt     # Complete documentation
Enter fullscreen mode Exit fullscreen mode

Think of it as robots.txt but for AI understanding, not crawling.

2. Schema JSON-LD Markup

Structured data that tells AI exactly what your product is:

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "Queryra",
  "applicationCategory": "WordPress Plugin",
  "operatingSystem": "WordPress/WooCommerce"
}
Enter fullscreen mode Exit fullscreen mode

3. AI Crawler Permissions

Explicitly allowing AI bots to crawl your site:

User-agent: GPTBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: ClaudeBot
Allow: /
Enter fullscreen mode Exit fullscreen mode

Many WordPress sites block these by default without realizing it.

4. Comparison Pages

A /compare page with structured competitor analysis gives AI the context it needs to position your product relative to alternatives.


What Actually Worked?

It's too early for definitive data, but here's what I've noticed after one week:

  • Perplexity (which does live web search) started surfacing Queryra in results faster than others
  • Schema markup alone isn't enough — AI needs text content around it
  • llms.txt is promising but adoption is still early
  • Comparison pages seem to give AI the most useful context for recommendations

The honest answer: none of these are a silver bullet yet. AI discoverability is where SEO was in 2005 — nobody really knows the rules, but the people who figure it out early will have a massive advantage.


Key Takeaways for Developers

If you're launching anything in the WordPress/WooCommerce ecosystem (or anywhere, really):

  1. Check your robots.txt — make sure you're not accidentally blocking GPTBot, ClaudeBot, and other AI crawlers
  2. Add structured data — Schema.org markup helps AI understand what your product actually is
  3. Create an llms.txt file — it's 10 minutes of work and the standard is growing
  4. Write comparison content — AI loves structured "X vs Y" context
  5. Don't rely on WordPress.org alone — being listed in the plugin directory doesn't mean AI knows about you

Try This Right Now

Ask ChatGPT or Claude:

"What are the best AI search plugins for WooCommerce?"

Then search for the same thing in the WordPress plugin directory. Compare the results. The gap is staggering.

That gap is an opportunity — for your products too.


I'm building Queryra — a semantic search plugin for WooCommerce. Free in the WordPress plugin directory (search "Queryra" in Plugins → Add New).


Has anyone else experimented with making their products visible to AI assistants? What worked for you?

Top comments (0)