Research Shows Differences In ChatGPT And Google AIO Answers via @sejournal, @martinibuster

New research from enterprise search marketing platform BrightEdge discovered differences in how Google and ChatGPT surface content. These differences matter to digital marketers and content creators because they show how content is recommended by each system. Recognizing the split enables brands to adapt their content strategies to stay relevant across both platforms.

BrightEdge’s findings were surfaced through an analysis of B2B technology, education, healthcare, and finance queries. It’s possible to cautiously extrapolate the findings to other niches where there could be divergences in how Google and ChatGPT respond, but that’s highly speculative, so this article won’t do that.

Core Differences: Task Vs. Information Orientation

BrightEdge’s research discovered that ChatGPT and Google AI Overviews take two different approaches to helping users take action. ChatGPT is more likely to recommend tools and apps, behaving in the role of a guide for making immediate decisions. Google provides informational content that encourages users to read before acting. This difference matters for SEO because it enables content creators and online stores to understand how their content is processed and presented to users of each system.

BrightEdge explains:

“In task-oriented prompts, ChatGPT overwhelmingly suggests tools and apps directly, while Google continues to link to informational content. While Google thrives as a research assistant, ChatGPT acts like a trusted coach for decision making, and that difference shapes which tool users instinctively choose for different needs.”

Divergence On Action-Oriented Queries

ChatGPT and Google tend to show similar kinds of results when users are querying for comparisons, but the results begin to diverge when the user intent implies they want to act. BrightEdge found that prompts about credit card comparisons or learning platforms generated similar kinds of results.

Questions with an action intent, like “how to create a budget” or “learn Python,” lead to different answers. ChatGPT appears to treat action intent prompts as requiring a response with tools, while Google treats them as requiring information.

BrightEdge notes that Healthcare has the highest rate of divergence:

“At 62% divergence, healthcare demonstrates the most significant split between platforms.

• When prompts pertain to symptoms or medical information, both ChatGPT and Google will mention the CDC and The Mayo Clinic.
• However, when prompted to help with things like “How to find a doctor,” ChatGPT pushes users towards Zocdoc, while Google points to hospital directories.”

B2B Technology niche has the second highest level of divergence:

“With 47% divergence, B2B tech shows substantial platform differences.

• When comparing technology, such as cloud platforms, both suggest AWS and Azure.
• When asked “How to deploy things (such as specific apps),” ChatGPT relies on tools like Kubernetes and the AWS CLI, while Google offers tutorials and Stack Overflow.”

Education follows closely behind B2B technology:

“At 45% divergence, education follows the same trend.

• When comparing “Best online learning platforms,” both platforms surface Coursera, EdX, and LinkedIn Learning.
• When a user’s prompt pertains to learning a skill such as “How to learn Python,” ChatGPT recommends Udemy, whereas Google directs users to user-generated content hubs like GitHub and Medium.”

Finance shows the lowest levels of divergence, at 39%.

BrightEdge concludes that this represents a “fundamental shift” in how AI platforms interpret intent, which means that marketers need to examine the intent behind the search results for each platform and make content strategy decisions based on that research.

Tools Versus Topics

BrightEdge uses the example of the prompt “What are some resources to help plan for retirement?” to show how Google and ChatGPT differ. ChatGPT offers calculators and tools that users can act on, while Google suggests topics for further reading.

Screenshot Of ChatGPT Responding With Financial Tools

There’s a clear difference in the search experience for users. Marketers, SEOs, and publishers should consider how to meet both types of expectations: practical, action-based responses from ChatGPT and informational content from Google.

Takeaways

• Split In User Intent Interpretation:
  Google interprets queries as requests for information, while ChatGPT tends to interpret many of the same queries as a call for action that’s solved by tools.
• Platform Roles:
  ChatGPT behaves like a decision-making coach, while Google acts as a research assistant.
• Domain-Specific Differences:
  Healthcare has the highest divergence (62%), especially in task-based queries like finding a doctor.
  B2B Technology (47%) and Education (45%) also show significant splits in how guidance is delivered.
  Finance shows the least divergence (39%) in how results are presented.
• Tools vs. Topics:
  ChatGPT recommends actionable resources; Google links to authoritative explainer content.
• SEO Insight:
  Content strategies must reflect each platform’s interpretation of intent. For example, creating actionable responses for ChatGPT and comprehensive informational content for Google. This may even mean creating and promoting a useful tool that can surface in ChatGPT.

BrightEdge’s research shows that, for some queries, Google and ChatGPT interpret the same user intent in profoundly different ways. While Google treats action-oriented queries as a prompt to deliver informational content, ChatGPT responds by recommending tools and services users can immediately act on. This divergence calls attention to the need to understand when ChatGPT is delivering actionable responses in order for marketers and content creators to create platform-specific content and web experiences.

Read the original research:

Brand Visibility: ChatGPT and Google AI Approaches by Industry

Featured Image by Shutterstock/wenich_mit

Google Explains The Process Of Indexing The Main Content via @sejournal, @martinibuster

Google’s Gary Illyes discussed the concept of “centerpiece content,” how they go about identifying it, and why soft 404s are the most critical error that gets in the way of indexing content. The context of the discussion was the recent Google Search Central Deep Dive event in Asia, as summarized by Kenichi Suzuki.

Main Body Content

According to Gary Illyes, Google goes to great lengths to identify the main content of a web page. The phrase “main content” will be familiar to those who have read Google’s Search Quality Rater Guidelines. The concept of “main content” is first introduced in Part 1 of the guidelines, in a section that teaches how to identify main content, which is followed by a description of main content quality.

The quality guidelines define main content (aka MC) as:

“Main Content is any part of the page that directly helps the page achieve its purpose. MC can be text, images, videos, page features (e.g., calculators, games), and it can be content created by website users, such as videos, reviews, articles, comments posted by users, etc. Tabs on some pages lead to even more information (e.g., customer reviews) and can sometimes be considered part of the MC.

The MC also includes the title at the top of the page (example). Descriptive MC titles allow users to make informed decisions about what pages to visit. Helpful titles summarize the MC on the page.”

Google’s Illyes referred to main content as the centerpiece content, saying that it is used for “ranking and retrieval.” The content in this section of a web page has greater weight than the content in the footer, header, and navigation areas (including sidebar navigation).

Suzuki summarized what Illyes said:

“Google’s systems heavily prioritize the “main content” (which he also calls the “centerpiece”) of a page for ranking and retrieval. Words and phrases located in this area carry significantly more weight than those in headers, footers, or navigation sidebars. To rank for important terms, you must ensure they are featured prominently within the main body of your page.”

Content Location Analysis To Identify Main Content

This part of Illyes’ presentation is important to get right. Gary Illyes said that Google analyzes the rendered web page to located the content so that it can assign the appropriate amount of weight to the words located in the main content.

This isn’t about the identifying the position of keywords in the page. It’s just about identifying the content within a web page.

Here’s what Suzuki transcribed:

“Google performs positional analysis on the rendered page to understand where content is located. It then uses this data to assign an importance score to the words (tokens) on the page. Moving a term from a low-importance area (like a sidebar) to the main content area will directly increase its weight and potential to rank.”

Insight: Semantic HTML is an excellent way to help Google identify the main content and the less important areas. Semantic HTML makes web pages less ambiguous because it uses HTML elements to identify the different areas of a web page, like the top header section, navigational areas, footers, and even to identify advertising and navigational elements that may be embedded within the main content area. This technical SEO process of making a web page less ambiguous is called disambiguation.

3. Tokenization Is Foundation Of Google’s Index

Because of the prevalence of AI technologies today, many SEOs are aware of the concept of tokenization. Google also uses tokenization to convert words and phrases into a machine-readable format for indexing. What gets stored in Google’s index isn’t the original HTML; it’s the tokenized representation of the content.

4. “Soft 404s Are A Critical Error

This part is important because it frames soft 404s as a critical error. Soft 404s are pages that should return a 404 response but instead return a 200 OK response. This can happen when an SEO or publisher redirects a missing web page to the home page in order to conserve their PageRank. Sometimes a missing web page will redirect to an error page that returns a 200 OK response, which is also incorrect.

Many SEOs mistakenly believe that the 404 response code is an error that needs fixing. A 404 is something that needs fixing only if the URL is broken and is supposed to point to a different URL that is live with actual content.

But in the case of a URL for a web page that is gone and is likely never returning because it has not been replaced by other content, a 404 response is the correct one. If the content has been replaced or superseded by another web page, then it’s proper in that case to redirect the old URL to the URL where the replacement content exists.

The point of all this is that, to Google, a soft 404 is a critical error. That means that SEOs who try to fix a non-error event like a 404 response by redirecting the URL to the home page are actually creating a critical error by doing so.

Suzuki noted what Illyes said:

“A page that returns a 200 OK status code but displays an error message or has very thin/empty main content is considered a “soft 404.” Google actively identifies and de-prioritizes these pages as they waste crawl budget and provide a poor user experience. Illyes shared that for years, Google’s own documentation page about soft 404s was flagged as a soft 404 by its own systems and couldn’t be indexed.”

Takeaways

• Main Content
  Google gives priority to the main content portion of a given web page. Although Gary Illyes didn’t mention it, it may be helpful to use semantic HTML to clearly outline what parts of the page are the main content and which parts are not.
• Google Tokenizes Content For Indexing
  Google’s use of tokenization enables semantic understanding of queries and content. The importance for SEO is that Google no longer relies heavily on exact-match keywords, which frees publishers and SEOs to focus on writing about topics (not keywords) from the point of view of how they are helpful to users.
• Soft 404s Are A Critical Error
  Soft 404s are commonly thought of as something to avoid, but they’re not generally understood as a critical error that can negatively impact the crawl budget. This elevates the importance of avoiding soft 404s.

Featured Image by Shutterstock/Krakenimages.com

Google’s Mueller Advises Testing Ecommerce Sites For Agentic AI via @sejournal, @martinibuster

Google’s John Mueller re-posted the results of an experiment that tested if ecommerce sites were accessible by AI Agents, commenting that it may be useful to check if your ecommerce site works for AI agents that are shopping on behalf of actual customers.

AI Agent Experiment On Ecommerce Sites

Malte Polzin posted commentary on LinkedIn on an experiment he did to test if the top 50 Swiss ecommerce sites were open for business for users who are shopping online with ChatGPT agents.

They reported that most of the ecommerce stores were accessible to ChatGPT’s AI agent but he also found some stores were not for a few reasons.

Reasons Why ChatGPT’s AI Agent Couldn’t Shop

• CAPTCHA prevented ChatGPT’s AI agent from shopping
• Blocked by Cloudflare’s Turnstile tool that’s a CAPTCHA alternative.
• Store blocked access with a maintenance page
• Bot defense blocked access

Google’s John Mueller Offers Advice

Google’s John Mueller recommended checking if your ecommerce store is open for business to shoppers who use AI agents. It may become more commonplace that users employ agentic search for online shopping.

He wrote:

“Pro tip: check your ecommerce site to see if it works for shoppers using the common agents. (Or, if you’d prefer they go elsewhere because you have too much business, maybe don’t.)

Bot-detection sometimes triggers on users with agents, and it can be annoying for them to get through. (Insert philosophical discussion on whether agents are more like bots or more like users, and whether it makes more sense to differentiate by actions rather than user-agent.)”

Should SEOs Add Agentic AI Testing To Site Audits?

SEOs want to consider adding Agentic AI accessibility to their site audits for ecommerce sites. There may be other use cases where an AI agent may need access to filling out forms, for example on a local services website.