In a significant evolution of conversational search, ChatGPT can now recommend products—and not as ads, but as part of its core search capabilities. For businesses looking to expand their digital reach and gain high-intent exposure, this presents a game-changing opportunity.
As part of OpenAI’s latest updates, ChatGPT is being used as a product discovery tool powered by real-time web data. If your business offers products or services online, it’s now possible—without paying for ads—to have them surfaced in natural, intent-driven interactions within ChatGPT.
Let’s unpack how this works, what it means for your business, and how Velocity can help you take advantage of it.
Understanding ChatGPT Product Recommendations
The Technology Behind It: OAI-SearchBot
How to Make Your Products Discoverable in ChatGPT
What This Means for Brands
Velocity’s Role: Helping You Get Discovered
Next Steps
FAQs
As ChatGPT continues to evolve beyond conversation into a powerful discovery tool, it’s now capable of recommending real products based on user queries. This shift opens new opportunities for businesses to appear directly in AI-powered search experiences—without relying on paid advertising.
When users ask ChatGPT questions such as:
“Best ergonomic chairs for under R2000”
“Affordable solar panel kits in South Africa”
“What’s a good CRM for SMEs?”
The model may return a curated list of relevant products pulled from the open web, based on product details, specifications, pricing, and availability.
These recommendations are not paid placements—they’re organic, meaning your product or service could appear simply by being discoverable and relevant.
To support this feature, OpenAI developed a specialised crawler called OAI-SearchBot. This bot actively scans the web for up-to-date product pages and related content. It indexes publicly available information such as:
Product titles and descriptions
Features and specifications
Prices and availability
Reviews and ratings
Structured metadata (e.g. Schema.org)
Your content needs to be accessible and well-structured to be included.
Important: OpenAI has also deployed another bot, GPTBot, which is used to train its general language models. While GPTBot helps shape how ChatGPT understands language, OAI-SearchBot is the one responsible for powering real-time product discovery.
OpenAI has outlined a few key requirements for your products to be eligible:
In your robots.txt file, ensure the following lines are included to permit access:
User-agent: GPTBot
Allow: /
User-agent: oai-searchbot
Allow: /
This ensures both bots can crawl your website without restriction. If you’re not familiar with how to access or update your robots.txt, Velocity can assist with a technical audit and implementation.
Implement Schema.org structured data for all your product and service pages. This allows crawlers to clearly identify the product name, price, availability, rating, and more.
Here’s an example of what this might look like:
{ "@context": "https://schema.org",
"@type": "Product",
"name": "Solar Battery Backup Kit",
"description": "A complete backup solution for off-grid energy systems.",
"brand": {
"@type": "Organization",
"name": "EcoVolt"
},
"offers": {
"@type": "Offer",
"price": "8999.00",
"priceCurrency": "ZAR",
"availability": "https://schema.org/InStock"
}
}
Proper metadata helps the search bot understand your offering and rank it accurately in response to user prompts.
While ChatGPT is not a traditional search engine, its logic is rooted in understanding user intent. To maximise visibility, your site should include content that directly addresses popular product-related questions, such as:
“Which marketing automation platform is best for start-ups?”
“Compare VoIP tools for remote teams”
“What’s the best laptop for designers in 2025?”
Use blog posts, landing pages, and comparison articles to create AI-friendly content that aligns with real user questions. The more relevant and structured your content is, the higher your chance of being pulled into a ChatGPT recommendation.
This feature introduces an entirely new traffic channel: AI-driven product discovery.
Free, organic exposure to potential customers via ChatGPT
High purchase intent from users actively seeking recommendations
Real-time accuracy driven by live page crawling
No reliance on paid media to reach discovery
For early adopters, this is a chance to appear alongside major retailers and brands—simply by optimising your content for accessibility and structure.
At Velocity, we specialise in helping brands adapt to shifts in the digital marketing landscape. Our technical and content teams are already working to help clients:
Configure their websites for crawler visibility
Implement structured data for all product and service pages
Write intent-driven articles and landing pages aligned with AI discovery
Track and optimise performance via UTM parameters and analytics
As OpenAI rolls out product feed submissions in future updates, we’ll be supporting our clients in structuring and maintaining these feeds to ensure maximum exposure in ChatGPT and other AI tools.
This is not just a new search feature—it’s a new way users find products and make decisions. Being early and strategic gives your brand an edge.
Here’s how to begin:
Audit your robots.txt and enable access to GPTBot and OAI-SearchBot
Apply Schema.org structured data to your product or service pages
Align your content strategy with real search intent
Monitor performance and refine your approach over time
Need help? Reach out to our team and let’s discuss how to futureproof your content for the age of AI discovery.
Generative Engine Optimisation refers to the practice of improving your content so that it is discoverable and recommended by AI tools like ChatGPT, Google Gemini, and Perplexity AI. Unlike traditional SEO, GEO focuses on optimising for natural language queries, structured data, and crawlability by AI bots rather than just ranking on Google.
To rank in ChatGPT’s native product responses, you need to:
Allow crawling by OpenAI’s oai-searchbot in your robots.txt
Add structured data using Schema.org on your product or service pages
Write content that matches real user intent (e.g. “Best tools for remote marketing teams”)
Keep your product data accurate and updated
There is no ad programme—recommendations are currently organic.
OpenAI uses two main crawlers:
GPTBot: Used for training large language models
oai-searchbot: Used to fetch live data, especially for product recommendations
Make sure these bots are allowed to crawl your content by adjusting your robots.txt file accordingly.
Yes. Google Gemini relies on Google’s standard web crawler, Googlebot, as well as other indexing systems within the Google Search ecosystem. If your site is already well optimised for SEO, much of that work applies to generative models like Gemini too.
Yes. Structured data (like Schema.org markup) helps AI systems understand your content in context. For product pages, this includes key information such as:
Product name
Price
Availability
Brand
Ratings
The clearer your markup, the better the AI’s ability to recommend your product accurately.
Absolutely. While the focus is on product discovery, well-structured service offerings with clear descriptions, use cases, pricing models, and reviews can also be surfaced in certain types of recommendations—particularly in B2B contexts.
Not yet. OpenAI is working on a product feed submission system, which will allow merchants to submit their product catalogue directly. In the meantime, your best strategy is to optimise content and structure for oai-searchbot crawling.
Use UTM tracking links on your product URLs, and monitor traffic sources in GA4. While OpenAI does not currently provide referrer data for ChatGPT, you may see direct or unlabelled traffic increases on high-performing product pages.
No—but it complements it. GEO (Generative Engine Optimisation) should be seen as an extension of your current SEO strategy. Traditional search engines still dominate traffic, but AI tools are quickly becoming a new layer of search discovery, especially for high-intent queries.
Not necessarily. However, you should:
Write more content that mirrors how people ask questions
Use comparison formats (e.g. “Top 5 CRMs for Start-ups”)
Regularly update product or service pages
Ensure pages are technically clean, fast, and crawlable