Late last year, Perplexity launched Shop like a Pro, letting users ask for product recommendations and buy directly from the platform.
Then earlier this year, OpenAI released Operator, an AI agent that can perform browser-based tasks for you—including buying things online.
Now, in April 2025, OpenAI strikes again with its ChatGPT Shopping feature. Similar to Perplexity’s Shop like a Pro or Google Shopping, ChatGPT now recommends products with clear AI-generated descriptions, summarized reviews, and direct links to websites where you can purchase them.
Like we shared in our previous article, AI is fundamentally reshaping how people shop. Ecommerce brands should take note and re-evaluate their acquisition strategies.
What is ChatGPT shopping?
Previously, when you searched for a product in ChatGPT, it simply listed recommendations in bullet points—with minimal context or reasoning behind each item.
The latest update introduces a much richer shopping experience. Now, if you enter a shopping-related prompt with Search turned on, you’ll receive product cards with:
- AI-generated product descriptions
- Summarized reviews
- A list of merchants and purchasing options
It feels like a blend of Google Shopping and a personal shopper rolled into one.
How does ChatGPT select products to recommend?
Much like Google Shopping, ChatGPT uses your query, any specific instructions you include, and your past interactions (if memory is on) to tailor recommendations. It learns from your preferences and improves over time.
You can also guide it explicitly by providing more context—or reset/delete memory to start fresh.
Here’s an example from ChatGPT:
If a user asks ChatGPT for help finding goofy costumes for their two large dogs, ChatGPT will consider general factors, such as price, customer ratings, and ease of use, as well as specific criteria provided by the user, like sizing and the desired costume vibe. If the user had previously indicated a dislike for clowns, the model might also consider that and leave out clown costumes.
When recommending products, ChatGPT considers these factors:
- Structured metadata from third-party providers (e.g., price, product description) and other third-party content (e.g., reviews)
- Model responses generated by ChatGPT before it considers any new search results
- OpenAI safety standards
According to Adam Fry, ChatGPT’s search product lead, the process isn't keyword- or signal-based like traditional search. Instead, it’s a more conversational, context-aware experience.
How does ChatGPT select merchants?
Each product card includes a list of merchants. And the order of these merchants is determined by merchant and product metadata sourced from third-party providers.
In terms of how it ranks the merchants, OpenAI clarifies:
“Currently, the order in which we display merchants is predominantly determined by these providers. We do not re-rank merchants based on factors such as price, shipping, or return policies.”
This likely means that the order came from Bing Shopping or Shopify listings, as ChatGPT doesn’t control or optimize the ranking of the product listings.
As of now, merchants can’t submit product feeds directly to OpenAI—but that’s reportedly in the works.
Our experience with ChatGPT Shopping
When we searched for “best noise-canceling headphones under $400,” ChatGPT returned a clean set of product cards. Each recommendation was labeled—for example: “Top pick,” “Best value,” or “Most comfortable.”
Interestingly, ChatGPT only showed the black version of the Sony headphones, even though other colors exist. This points to potential limitations in the product metadata.
Clicking on a product card reveals:
- AI-generated summaries of why you might like the item
- Star ratings and reviews (from third-party sources)
- A list of purchasing options, with prices and direct links
Currently, these are non-sponsored results—meaning brands can’t pay to be featured (yet, maybe).
The order of purchasing options appears to be sorted by price, but OpenAI claims they don’t re-rank them.
A quick search on Bing didn’t surface some of the same merchants, like The Teds Store, 6ave Electronics or The Pixel Hub—which sounds contradictory. But that could be just be a coincidence.
According to OpenAI, ratings are taken from third party providers and may be aggregated. They’re also not verified by OpenAI themselves.
Each card also includes what looks like an unbiased view listing positive and negative points, linking to forums, reviews, and authoritative sites like Reddit or The Verge, depending on your query.
You can also ask follow-up questions about specific products by clicking on it, revealing the “Ask about this” button.
This is how ChatGPT responded. What’s useful is that it also shows images and videos. What’s not useful is that it doesn’t tell me where I can buy them.
How is ChatGPT Shopping different from Perplexity’s Shop like a Pro?
While both ChatGPT Shopping and Perplexity Shop like a Pro offer product cards with AI-generated reviews and purchase links, they diverge in important ways:
Visual search
Perplexity has Snap to Shop, allowing you to upload an image and find where to buy it—like Google Lens.
ChatGPT doesn’t support this yet. Funnily, when I tried to upload an image and ask where to buy it, it tries to confuse me with this:
Direct checkout function
Perplexity partners with merchants to enable one-click checkouts with perks like free shipping.
ChatGPT only offers external links—you’ll need to go to the merchant site to complete your purchase. No free shipping either.
Key features
Perplexity includes bullet-point summaries of key product features. ChatGPT focuses more on personalized AI summaries, but may miss structured info that helps comparison shoppers.
Memory and context
Perplexity’s recommendations are prompt-based only. ChatGPT, however, uses memory to tailor results to your preferences and past activity.
That makes the recommended products much more personalized and relevant.
That said, it can be a double-edged sword—one-off interactions you didn’t mean to influence recommendations might still affect them. You can, however, instruct ChatGPT to ignore those.
Paid placements
Neither platform supports sponsored listings (yet), making them more organic than Google Shopping or Amazon.
How can ecommerce brands gain visibility on ChatGPT Shopping?
When Perplexity launched Shop like a Pro, it signaled how AI would disrupt product discovery. ChatGPT’s entry confirms that change is here.
Regardless, a lot of the same principles still apply—only now, they matter more than ever.
AI is changing how people discover and decide on products, but the foundation for visibility remains the same: rich, accurate, and relevant product content. What’s changing is who or what is reading that content. It’s no longer just humans—it’s also AI models that interpret your data and decide whether or not your product appears in a recommendation.
That’s why it’s critical for ecommerce brands to:
Invest in high-quality metadata and structured content
AI models like ChatGPT and Perplexity don’t just “read” your page—they interpret structured data like schema markup, product attributes, availability, and pricing from third-party feeds and sources.
Incomplete or inconsistent metadata weakens your discoverability, making it harder for your products to appear in relevant searches—even if they’re a perfect fit.
Therefore, it’s important to optimize your metadata.
Hypotenuse AI helps ecommerce teams enrich and structure product data at scale—filling in gaps, standardizing information, and applying consistent formatting across thousands of SKUs.
Write product descriptions that speak to real customer intent
AI models don’t just match keywords—they aim to understand what the shopper is looking for: What problem are they trying to solve? What features matter most in their context? What benefits are they weighing?
Today’s shoppers—and the AI models assisting them—look for content that reflects real intent and use cases: Is this jacket warm enough for winter hikes? Will this standing desk fit in a small home office?
Hypotenuse AI helps ecommerce brands generate product descriptions that answer those questions directly.
It turns raw product specs into compelling narratives that highlight benefits, solve shopper pain points, and adapt to different contexts—whether you're selling the same product to parents, professionals, or outdoor enthusiasts.
Ensure consistency across your channels
Your brand’s product data appears in more places than just your website—think marketplaces, resellers, product feeds, and more.
If those channels present mismatched specs, inconsistent naming, or different formats, AI models will have a harder time confidently surfacing your products.
With Hypotenuse AI, brands can centralize and synchronize product data and content across multiple platforms.
That means formatting specifications correctly, aligning product messaging, and preserving factual consistency—no matter where a product appears.
Strengthen your digital footprint and review presence
ChatGPT doesn’t just rely on your product feed—it pulls from forums, third-party reviews, publisher content, and other signals across the web. If your brand isn’t showing up in those places, it’s less likely to be recommended.
Encourage authentic customer reviews on platforms your audience trusts. Partner with credible publishers and creators. Seed relevant conversations in forums and community spaces where people research and compare products.
The more consistently your brand appears in high-quality, trusted sources, the stronger your presence in AI-powered shopping results.
The takeaway
Product discovery has been moving toward intent and context for a while—AI is just accelerating it.
With ChatGPT and Perplexity now recommending products directly, your content needs to do more than sit on a page. It has to be structured, consistent, and actually useful—because AI is the one deciding what shows up.
The fundamentals haven’t changed. But how well you execute them matters more than ever.
