
OpenAI is actively testing new search result formats inside ChatGPT, according to Search Engine Roundtable, and the variations are significant enough that SEOs and content strategists need to pay attention now. The tests include web-only result pages, a collapsed AI answer with a 'See More' expansion button, and a hybrid layout that shows web results first before surfacing ChatGPT's standard AI response. If you've been thinking about AI visibility tracking as a nice-to-have, these experiments suggest it's quickly becoming essential.
Barry Schwartz of Search Engine Roundtable reported the findings on November 16, 2026, noting that he could not replicate any of the tests himself. The variations were each spotted independently by different search marketers on X, suggesting OpenAI is running multiple parallel interface experiments rather than a single staged rollout.
What OpenAI Is Testing Inside ChatGPT Search
According to Search Engine Roundtable's report, there are at least three distinct format variations being tested right now. Each one points to a different philosophy about how an AI assistant should present search results.
The first variation, spotted by Joe Youngblood on X, shows ChatGPT returning web-only results with no AI-generated answer layered on top. Schwartz noted he thought he had seen this before and dismissed it as a bug. With multiple tests surfacing in the same week, he decided it was worth covering.
The second variation, spotted by Juozas Kaziukenas on X, flips the order: ChatGPT shows web results first and then follows with its standard, more polished AI response. That's the reverse of what most users see today, where the AI answer dominates and web links are secondary.
The third variation, spotted by Fernando on X, puts the AI answer at the top but collapses it behind a 'See More' button before listing web results below. Schwartz drew a direct comparison to how Google handles AI Overviews: a summary up front, expandable for detail, with traditional links underneath.
What These ChatGPT Search Experiments Mean for SEO Strategy
Here's what I find most interesting about these tests: OpenAI isn't picking one format and rolling it out. They're running several very different philosophies at the same time. That tells you they haven't settled on what a 'ChatGPT search result' should even look like yet.
And that uncertainty has real consequences for your content strategy. If the web-only results format wins, traditional link-building and on-page SEO matter more than ever inside ChatGPT. If the collapsed 'See More' format wins, you're essentially looking at Google's AI Overviews model inside a competing product. Your content needs to be citation-worthy AND clickable.
The hybrid layout, where web results appear before the AI answer, is the most disruptive scenario for established sites. It would mean ChatGPT is leaning into acting like a traditional search engine for some queries, which opens the door for ranking signals we already understand to carry weight inside the platform.
What's consistent across all three formats? Web results show up. That means crawlability, indexation, and structured data are not going away. If your site is hard to crawl or your pages lack clear entity signals, you're invisible in every version of this test.
ChatGPT Search Formats and the Parallel to Google AI Overviews
The 'See More' format is the one that should grab your attention most if you've been following Google's AI Overview rollout. Google spent years iterating on how to show an AI answer without killing click-through. The collapsed summary with expandable detail is the model they landed on. If OpenAI is testing the same pattern, they're converging on a similar user experience philosophy.
That convergence matters for how you write content. Concise, well-structured answers that can be pulled and summarized tend to get cited in AI Overviews. The same logic is likely to apply inside ChatGPT if this format wins out. Clear headers, tight paragraphs, direct answers near the top of the page: these aren't just SEO best practices anymore. They're your ticket into AI-generated summaries across multiple platforms.
How Agencies and Marketers Should Respond to ChatGPT's Search Format Tests
You don't need to overhaul your entire strategy this week. But you do need to stop treating ChatGPT as a purely generative tool that's separate from the search landscape. These tests confirm it's moving toward a search product, and your visibility inside it will increasingly depend on the same fundamentals you apply to Google.
- Audit your crawlability and indexation now. If ChatGPT's web results format indexes via Bing's index (which OpenAI has historically used), your Bing presence matters more than most agencies have given it credit for. Run a Bing Webmaster Tools check and make sure your pages are being crawled and indexed cleanly.
- Structure your content for citation. Regardless of which format wins, every version of this test surfaces web content. Write direct answers near the top of your pages, use descriptive H2s and H3s, and make sure your most important pages have clean schema markup.
- Track your AI visibility alongside traditional rankings. Knowing you rank on page one of Google is not the same as knowing whether you're being cited in ChatGPT, Perplexity, or AI Overviews. These are different surfaces with different citation logic.
- Watch the 'See More' format closely. If it rolls out broadly, your meta descriptions and page summaries become more important than they've been in years. That collapsed AI answer has to source from somewhere.
- Brief your clients proactively. If you run an agency, your clients are going to see screenshots of these tests circulating on social. Get ahead of the question by framing what it means and what you're already doing about it.
ChatGPT Search Tests in the Context of the Broader AI Search Race
This isn't happening in a vacuum. Google has been iterating on AI Overviews since the SGE era. Perplexity built its entire product on the blend-of-AI-and-links model. Microsoft has been embedding Copilot into Bing results for over a year. OpenAI is now testing the same blended format, and they're doing it in a product that already has enormous user trust for generative tasks.
I've watched the search industry spend two years arguing about whether AI would replace traditional results. What we're actually seeing is that AI and traditional results are merging. The question was never 'AI or links.' It's always been 'how do you present both in a way users trust?'
These ChatGPT tests suggest OpenAI is still working out that answer. And honestly, so is everyone else. The fact that they're running multiple parallel experiments tells you there's no settled UX consensus yet, even inside the company that built the most widely used AI product on the planet.
For the SEO community, the right posture is flexible fundamentals. Sites that are technically sound, semantically clear, and genuinely useful will show up in every version of this test. Sites built around loopholes and thin content will get squeezed no matter which format wins. That part hasn't changed.
If you want to keep a close eye on where your brand and content are showing up across AI search surfaces, Aergos has AI visibility tracking built into its dashboard so you're not flying blind as these formats evolve.
Frequently Asked Questions
Is your site set up to show up in search and AI?
Run a free instant audit. We check the on-page SEO basics and the files that control whether search engines and AI crawlers can access your site: robots.txt, sitemap.xml and llms.txt.
Free · instant · no sign-up for your first audit.
Related Articles
Glossary terms in this article
Brush up on the definitions.
Microsoft's free platform for monitoring a site's Bing search performance, submitting sitemaps, and diagnosing indexing issues on Bing and Yahoo.
The planning, development, and management of content to achieve specific business goals across all channels and formats.
A standardised format for providing information about a page and classifying its content so search engines can better understand it.
A software system that indexes web content and returns ranked results in response to user queries, including Google, Bing, Yahoo, DuckDuckGo, and AI-powered answer engines.
The extent to which a brand's content is referenced, cited, or surfaced in AI-generated answers from tools like ChatGPT, Gemini, and Perplexity.
Structured data code added to a page's HTML that helps search engines understand its content and enables rich results in SERPs.

About Matt Weitzman
Senior SEO/GEO Strategist & Founder
Matt has over 15 years of experience in technical SEO and digital marketing. He specializes in algorithmic recovery, enterprise architecture, and leveraging AI for content scaling. He is a frequent speaker at search marketing conferences.
More articles by Matt Weitzman

