ProductKnowledgeGraph is now available as a ChatGPT Plugin

ProductKnowledgeGraph is now available as a ChatGPT Plugin

When buyers ask AI what to buy, does it show your product — or your competitor?

More buying decisions are starting before a customer ever reaches a brand’s website.

A buyer can ask ChatGPT:

“What should I buy for this use case?”

“Which product is best for me?”

“Will this model work with what I already own?”

“Which brand would you recommend?”

“Where can I buy it?”

And increasingly, AI can help narrow the options before the buyer visits a product page.

That creates a new problem for brands and retailers.

Your product may be the right fit — but AI may not surface it.

It may recommend a competitor instead.

It may misunderstand an important product fact, miss a variant or compatibility detail, rely on outdated information, or fail to send the buyer to the correct official product page.

And if the buyer never reaches your website, traditional ecommerce analytics may never show you the opportunity you lost.

That is the problem ProductKnowledgeGraph is built to address.

Today, we’re making it easier to work with that problem directly inside ChatGPT.

ProductKnowledgeGraph is now available in ChatGPT

We’ve launched the Product Knowledge Graph GPT Plugin, allowing authorized ProductKnowledgeGraph customers to work with their product intelligence directly from a ChatGPT conversation.

Instead of moving between dashboards to investigate individual questions, you can now simply ask.

For example:

• Are buyers finding my products when they ask AI what to buy — or am I invisible?
• Is AI recommending my products or my competitors — and why?
• Show me where I could be losing sales inside AI before buyers ever reach my website.

ProductKnowledgeGraph can then work with the authorized PKG workspace behind the conversation.

From product data to buyer questions

The important question is not simply whether an AI system “knows” that your product exists.

The commercial question is what happens when a real buyer asks AI for help choosing.

With the new ChatGPT Plugin, teams can use natural-language conversations to work through the ProductKnowledgeGraph workflow: inspect governed Product Truth, prepare buyer questions, run controlled AI Buyer Checks, identify products that were not surfaced, see which competitor products appeared instead, review product-information gaps, and compare results after approved corrections.

That makes a workflow that would normally require navigating multiple product-data and AI-observation screens accessible through questions asked directly in ChatGPT.

Find where the buyer is being lost

Consider a simple scenario.

You sell a product that is genuinely appropriate for a particular use case.

A prospective customer asks AI:

“What are the best products for what I need?”

AI recommends three competitors.

Your product never appears.

The customer may never search for your brand, never visit your website and never enter your ecommerce funnel.

For the business, that creates a blind spot.

ProductKnowledgeGraph is designed to help investigate that moment:

• What did the buyer ask?
• Which products did AI surface?
• Was our product present?
• Which competitors appeared instead?
• Is there an addressable product-information problem?
• What happens when we correct it and ask the same buyer question again?

Now those questions can be explored directly through ChatGPT.

Fix it. Then ask the same question again.

Finding a problem isn’t enough.

One of the central ideas behind ProductKnowledgeGraph is to connect observation with correction and retesting.

If a product was not surfaced because important authoritative product information was incomplete, unclear or inconsistent, the team can work through the governed correction process.

Then the same buyer question can be tested again.

That creates a much more useful loop:

And the new GPT Plugin makes that workflow substantially easier to access.

We don’t call an AI mention a sale

There is an important distinction behind ProductKnowledgeGraph.

A product being mentioned by AI does not prove that someone clicked.

A recommendation does not prove a purchase.

And a better result after changing product information does not, by itself, prove that the change caused additional revenue.

ProductKnowledgeGraph keeps those stages separate.

We want businesses to be able to distinguish between a product being available to AI, understood correctly, surfaced, recommended, connected to the right official destination, clicked and ultimately purchased.

That evidence matters as AI becomes a larger part of product discovery.

Your customers are already asking AI what to buy

The question for product companies is no longer simply whether their website ranks or their product feed is complete.

There is now another moment in the buying journey:

the conversation that happens before the website visit.

If AI understands your competitor better than it understands you, the buying decision may already be moving in the wrong direction before your website has a chance to convert the customer.

ProductKnowledgeGraph exists to make that part of the journey observable and actionable.

And now you can work with it directly from ChatGPT.

Try Product Knowledge Graph GPT

Connect your authorized ProductKnowledgeGraph workspace and start with one simple question: