Hey there.  Scott Austin here.

Today, we are continuing our dive into Agentic Commerce.  Shopify recently released their agentic channels for OpenAI and Microsoft CoPilot to most stores.  You can see it in your store at Settings > Sales channels and then Agentic storefronts on the bottom of the Sales channels page. 

Agentic Commerce

Let's spend a little time here explaining the term Agentic Commerce as it is a shift in the shopping experience. Instead of a purely human-centric model where shoppers rely on visually navigating a website and clicking through menus, natural language AI agents—such as ChatGPT, Gemini, Perplexity, or Google's AI-powered search—act as proactive, high-knowledge sales associates reducing the amount of effort required of customers to match products to their needs.

Here is how this evolving channel works:

AI as the Intermediary - Shoppers use conversational AI to conduct research, compare products, and ask complex logistical questions. For example, a customer might ask an AI, "What is the return policy for this brand's boots?" and follow up with, "Are they true to size?" The AI acts as the salesperson, synthesizing facts and providing natural language recommendations directly to the consumer. 

The bypassing of the Storefront - Because a bulk of brand discovery and education now happens in a chat interface before the customer ever visits your store, the traditional visual storefront becomes more of a fallback or secondary destination. In this model, platforms like Shopify function as massive data feeds supplying information to a global network of AI agents from all of their aggregated shops. 

Seamless In-Chat Checkout - The ultimate goal of the AI agent is to eliminate friction. As the technology matures, AI agents are increasingly capable of guiding users directly to a personalized checkout link or even completing transactions entirely within the chat environment, completely bypassing the traditional website navigation process.  Shopify and OpenAI are currently working on this.

The Shift to Generative Engine Optimization or GEO - For Agentic Commerce to function accurately, AI models require structured, verified data rather than traditional keyword-stuffed marketing copy. Traditional SEO focused on getting humans to click links on a search engine results page; GEO focuses on getting AI models to confidently cite your products and policies. 

The Black Box Problem - To understand why GEO is so critical, we have to talk about the Black Box Problem. If you have looked at your Shopify analytics lately, you might have started seeing platforms like chatgpt.com showing up as a referrer. It is awesome to know that AI is recommending your products, but here is the frustrating part: unlike traditional Google Search where you can see the exact search query a customer typed in, AI referrals are a total black box. Imagine a customer asks an AI three specific questions about your return policy and sizing, and only then decides to click the link to your product page. Your analytics will only show you that final click. You lose complete visibility into those high-intent questions that actually drove the purchase.

Shopify Knowledge Base App

If an AI is going to confidently guide a customer through a purchase, it needs certainty about a store's products, pricing, and policies. In the last podcast episode, I talked about how you can structure your product data for an agentic world.  But we also need to structure data around our store's policies.  To prepare for this, Shopify quietly released a free app called the Shopify Knowledge Base.  This is not a customer-facing FAQ page that sits on your website. Instead, it is a data feed of structured, verified facts about your brand that informs AI shopping agents.

Here is the breakdown of how the app operates and how your data is fed into AI agents.  This stuff is a bit technical, but I find it helps to understand what is happening under the hood.

When installed, the app automatically scans your store's backend settings to generate a baseline of verified "store facts." It pulls data directly from your store configuration, specifically analyzing your language settings, customer account options, shipping and delivery zones, and established return rules.

You can review these automated facts and override them with corrections, or manually add your own custom 1-to-2 sentence FAQs.  The app does not publish this text to a visual webpage. Instead, all of these FAQs are saved as metaobjects within your Shopify admin architecture. This is a critical step because it translates unstructured, human-readable website text into a highly structured, machine-readable format that AI models prefer.

Note: You may also hear about a related format called an llms.txt file. This is a separate, standardized text file acting somewhat like a "robots.txt" specifically for AI crawlers to summarize your site's content, but the Shopify Knowledge Base app natively formats your internal data as metaobjects.

All of this structured data is stored directly within your Shopify admin architecture. You do not need to host an external database or create a new public-facing page on your website. If you want to see exactly where this data lives, you can navigate to your Shopify admin dashboard, click on Content > Metaobjects > Standard product attributes.  You'll see a Definition named 'Question and Answer Pairs' that is added by the Knowledge Base app.  Each FAQ is its own entry.

How do AI agents find it? This is where the major shift in e-commerce technology happens. Traditionally, a search engine like Google would send a web crawler to "read" and scrape your visual storefront to guess your policies. With Agentic Commerce, the discovery process is much more direct.

Rather than relying on external web scrapers, Shopify acts as a direct, data feed. Shopify feeds your verified metaobjects directly into external AI platforms (like OpenAI) and its own proprietary Shopify Agent Platform. When a shopper asks ChatGPT or another AI agent a question about your brand, the AI doesn't have to go search your public website; it simply taps directly into this structured data feed to retrieve the verified facts you saved in your Knowledge Base.

This reduces AI hallucinations and increases accuracy. 

And there's an additional value of this app that goes way beyond just storing facts. The secret weapon here is the FAQ Query Log. As customers interact with AI agents, the app actually tracks what they are asking. The golden metric you need to watch is the "Unanswered queries." I want to emphasize why this data is pure gold. These unanswered questions are a direct window into your customer friction points and content gaps. If shoppers are repeatedly asking an AI if your boots run true to size or what your exact exchange policy is, and the AI can't find the answer, you are likely losing the sale. It’s not just about feeding the AI; it’s user journey intelligence. It's the solution to the black box problem we discussed earlier.

Setting up the app

Let's get into the step-by-step for setting up the app.

Step 1 is Installation. You can add the app from the Shopify app store.  There's a link to the listing in the show notes. 

Step 2 is Audit your automated FAQs. When you first open the app, you'll see that Shopify's AI has already created a number of FAQs for your store.  You can see them at the bottom of the screen in the section called Store FAQs.  The ones under the Default tab are the one generated by Shopify by looking at the content and policies for your store.  You should go through entries and review the facts Shopify generated. If an automated fact is wrong, you simply click the Override button and manually correct the answer.

Step 3 is to Close the Unanswered gap. Go to the Top unanswered questions section in the app. Hover over a popular query the AI could not answer, and click Answer. This will create a FAQ that the Shopify AI could not automatically generate.  Once created, these FAQs show in the Store FAQs section under the Custom tab.

Step 4 is to monitor your FAQs.  Over time, the app will compile a FAQ query log and list out more unanswered questions for you to answer.

Testing the App

You can test your setup. The app has a built-in Test your knowledge base feature. You can type in test queries, which are flagged with a special Test badge so they do not mess up your data, to preview exactly which resources the AI is matching to formulate its answer.

What Should Stores Do Now?

As a Shopify store owner, you wear a lot of hats.  You can't complete all of the tasks on your to-do list.  So let's be realistic on the time investment you should be making in agentic.  Here's my simple advise for you.  

Structuring your product data is the first step in preparing your store.  I covered that in Episode 182 of this podcast.  Our own Datify app will help you greatly with that.  We built that app because we saw the structured data need that wasn't being filled yet.  Our app is still in beta and we're making better every single week.  Use it to get your product categories set and then category metafields in place.

After that, install the Shopify Knowledge Base app and let create your default FAQs.

Now you'll have your basics in place.  Next is to watch the data.  There's alot of unknowns in this space and we're all learning together.  Monitor how agentic is affected your sales.  If this channel is working for you, then continue investing in it.  If it is not working for but is working for other brands, then its time to rethink your implementation.  

There's a lot of hype and interest around agentic.  But only time will show if the promise of agentic as a sales channel is real or hype.  So stay tuned for more from me on this in the future.

Thanks for listening.