In late November, OpenAI launched a new “shopping research” feature inside ChatGPT. If, as The Economist reported, two thirds of consumers in rich countries plan to use AI to help with holiday shopping, the release date is not a coincidence.
ChatGPT’s new functionality produces a tailored-for-you buying guide around a product category and is meant to curate pre-purchase research. I played around with it and found it incredibly helpful for the purchase of a higher-priced item I’m looking into.
For consumer-facing brands there are big implications around this new feature, because ChatGPT is both a behemoth and rapidly becoming the leader in terms of integrating shopping features into AI, which is quickly becoming the first stop on many purchase journeys.
What shopping research actually does
It’s pretty simple. Users describe what they need (for example, “I want a quiet vacuum for a small apartment” or “I am searching for a gaming laptop under $1,000”), and ChatGPT creates a personalized, structured buying guide. It pulls prices, specs, reviews, and availability from trusted retail and content sites, returning a ranked, conversational brief in minutes. People can click “More like this” or “Not interested,” and the recommendations update in real time based on that feedback.
In practice, a big chunk of comparison shopping that used to happen across dozens of browser tabs now happens inside one AI surface. I found the experience super efficient.
700M weekly users
Why should consumer tech companies care about this? The obvious answer is scale. ChatGPT is on track to reach roughly 700 million weekly active users in 2025, up sharply year-over-year. It is now in the 5 most visited websites on the internet. That means any new capability, including shopping research, is immediately exposed to usage at a scale most commerce and media channels never reach.
Why this is a PR issue, not just an ecommerce one
Shopping research is powered by an LLM trained specifically for shopping tasks and tuned to “read” the open web, product pages, reviews, editorial content, and more. The system is designed to reward structured, trustworthy, and consistent information about your products and brand.
That means:
- Your messaging is being interpreted first by an AI, not humans. If your narrative is fuzzy, fragmented, or contradictory across channels, the model’s summary of your products will be too.
- Clear positioning and proof points in owned content, FAQs, and executive commentary directly influence how ChatGPT describes your category, your differentiators, and your fit for a given user’s purchase decision.
Here’s a real example: A client’s product was being described by ChatGPT as “budget-friendly” when their positioning was actually “premium quality at fair prices.” The culprit? Inconsistent messaging across retailer pages and old press releases still ranking high in search.
As we have long said, every organization has a new ideal customer (bots) who love consistency. PR plays a critical role in how you cater to bots.
Editorial is AI Fuel
AI shopping and research tools lean heavily on “trusted” third-party content when deciding what to surface. Our own client research, done via Scrunch, shows that, in aggregate, ChatGPT relies more on editorial sources than the other major LLMs (in aggregate AI engines cite journalistic sources in 25 to 50 percent of answers).
This raises the bar for media relations. It means thin, generic mentions are less likely to matter, while detailed and differentiated coverage is more likely to be picked up in AI-generated recommendations.
Inconsistencies between what appears on your site and what appears in coverage can confuse the model and reduce your visibility in results. Signal clarity is the new gold.
Generative Engine Optimization and PR
To adapt, communications teams have to treat ChatGPT as a new, always-on editorial layer and build a dedicated GEO track alongside classic PR work.
Key moves include:
- Content hygiene: Audit product pages, boilerplates, and FAQs for clarity, specificity, and consistency on core claims—who it’s for, what it does best, and what it trades off. Technical specifics: ensure schema markup is complete, product descriptions follow a consistent structure, and key differentiators appear in the first 100 words of any description.
- PR targeting: Prioritize outlets ChatGPT scrapes and formats that provide structured, comparative, and data-backed coverage the model can easily parse and trust. Axios is a great example—their bullet-point structure, clear headers, and “Why it matters” format make content highly digestible for LLMs. Target reviews that compare multiple products side-by-side with specific criteria.
- Category narrative: Proactively define your category and its language so that, when people ask ChatGPT what to consider, your framing shows up in the explanation. If you sell “sleep headphones,” make sure that term appears consistently, as opposed to say, “bedtime audio” or “nighttime earbuds.”
What This Means
OpenAI is making the decision about what LLMs to optimize for much easier. With 700 million weekly users and growing commercial integrations, ChatGPT is becoming a primary interface between consumers and purchase decisions.
That means treating AI visibility as a fundamental channel, not an one-off experiment.
The window to establish your narrative in these systems is open, but it won’t stay that way. As more brands wake up to GEO, the cost and complexity of changing how AI describes you will only increase.
We can help you navigate the new, wild world that is blending PR, AI search content and other channels. Get in touch