Recently, I stumbled on an article that made a bold claim: GPT-5 has made SEO irreplaceable.

The argument? GPT-5 isn’t designed to “know everything” like a trivia champion. it’s designed to reason well and fetch facts from the web when it needs them.

That means the content it finds online (through Google/Bing) directly shapes the answers it gives. I wanted to put this to test and find out if this claim was really true.

Naturally, I started at a place close to home:

If GPT-5 really relies more on web searches, what does that mean for SaaS product discovery?

Because here’s the truth:
For 70% of SaaS businesses, pipeline metrics still come down to one thing:
People searching for a product online.

We’ve spent years running ads, optimizing keywords, and now—optimizing for AI search engines. But if GPT is building their systems to incorporate our existing efforts. That's a sweet deal.

The Test: Asking the same product searches on GPT 4o vs GPT 5

I pulled ~1,000 direct product search queries from Radix’s dataset spread across 5 key verticals.
Then I compared real-time GPT-5 answers with results we had from GPT-4o from a week ago.

These weren’t hypothetical prompts, this was actual, live product search data from across 6 SaaS categories.

How has GPT 5 changed AI visibility trends?

Tl;dr - The early movers are wining, GPT5 = More brand mentions and more citations.

Web searches are up — and SEO matters more than ever

• GPT-4o: Web search triggered in 29% of queries.

• GPT-5: Now 37%.

The LLM is able to easily recognise knowledge it doesn't have and search the right places to equip itself. These searches pull from Google/Bing rankings, meaning if you’re not ranking well, GPT-5 is far less likely to mention you.
For SaaS marketers, organic SEO now directly translates to AI search visibility.

More brands ranked per prompt — Less concentration at the top

Two trends stood out:

• Unique brands per category skyrocketed.

  • Example: Crypto Payments category went from 139 → 254 brands mentioned.

• Top 5,10 & 20 brands covering answers all saw drops, newer brands are recommended based on the context of the question.

Translation: GPT-5 spreads its recommendations wider. If you’re a challenger brand, this is your opening to get into the conversation.

I wanted to verify if this was just a correlation to more web search being triggered ( GPT recommends more brands when web search is on). But the truth is, GPT5 recommends more products especially when just relying on training data while the number of brands per answer remained the same when using a web search.

Not every vertical changed.

In mature markets (e.g., Subscription Billing Platforms), the same dominant brands kept their stronghold, unique brand mentions barely shifted. In emerging markets like AI visibility tools and AI sales tech, the diversity was huge.

Citations have increased — more diverse than ever before

There are more unique domains & URLs per answer. On average each category cited 200 different domains.

• Wikipedia & glossary sites are getting more traction.

• Reddit citations dropped from 11% → 7%.

• Comparison/Listicle pages are still the most influential content for purchase intent queries.

• Microsites and PR have gained even more traction. Influencer marketing is on the rise.

This could also hint that companies are optimising for AI search and specifically for ChatGPT citations and GPT-5 is indexing them stronger than before.

Reddit Is down

Total number of Reddit citations dropped from 11% → 7%. But it continues to influence ~85% of questions asked on ChatGPT.

This could mean GPT-5 is leaning more on structured, high-authority sources over open discussion forums for product recommendations. However, Reddit also continued to be the absolute most cited domain on ChatGPT.

What This Means for SaaS Marketers

• SEO is now Search Everywhere Optimisation. Ranking in Google/Bing = visibility in GPT-5 answers.

• Long-tail opportunity: Wider brand coverage means you can compete even if you’re not #1.

• Content format matters: Invest in glossaries, listicles, and high-authority original content.

• Emerging categories are gold mines: Position yourself strongly for niche usecases. GPT5 personalises answers far more than before.

• Prepare for “conversational competition”: GPT-5 will weigh your differentiators against more brands in every answer.

Getting Started with GEO

If GPT-5 has taught us anything, it’s that AI visibility is now tied directly to your SEO game, but winning here isn’t about guessing what might work.

The first step? Analyze what’s already working in your industry. This is exactly where Radix comes in.
Radix tracks AI search results across ChatGPT, Perplexity, and Google’s AI Overviews, showing you:

• Which prompts your brand appears for

• Which formats, citations, and domains dominate in your space

• Where you can break in before AI models “lock in” their favorites

Once you know the lay of the land, you can optimize the right content, target the right prompts, and claim your share of AI visibility before your competitors do.