Strategic Planning
Google just switched on agentic shopping. What that really means for brands
TechCrunch covered Google’s latest AI shopping update: conversational shopping in Search, new shopping flows in Gemini, agentic checkout, and an AI that literally phones local stores to check stock for you.
16 November 2025
9 min read
TechCrunch
What Google has actually shipped
From the outside it looks like a bundle of features. Under the hood it is a single idea: let an AI agent handle more of the journey from “I need something” to “order placed”.
According to TechCrunch and others, the update includes:
Conversational shopping in AI Mode
You describe what you want in natural language and refine it with follow up prompts. Google Search AI Mode turns that into a short list of products, with comparisons and historic pricing data pulled from a Shopping Graph of around 50 billion listings.
Agentic checkout (“Buy for me”)
You pick a product, set the options and a target price. Google tracks the price and, when it hits your number, the agent takes the product to the retailer’s checkout and uses your stored Google Pay details to complete the purchase (with a confirmation step). This starts with partners like Wayfair, Chewy, Quince and selected Shopify merchants.
“Let Google Call” for local stock and deals
You tell Google what you are looking for locally. An AI caller phones around nearby stores, checks availability and promotions, then sends you a neat summary with options.
If you step back, that is discovery, comparison, inventory and checkout all stitched together by an AI layer that is increasingly doing the clicking and dialling on your behalf.
Meanwhile in Redmond: Copilot is not far behind
This is not happening in a vacuum.
Microsoft has already rolled out Copilot Shopping, which helps users discover products, compare prices, track drops and make purchases directly in the Copilot experience.
On the merchant side, Microsoft has launched a Copilot Merchant Program, designed to let brands feed product data into Copilot so their catalogue can surface in AI powered shopping queries. Microsoft
There are also tests of native checkout and Shopify style integrations so that purchases can be completed inside Copilot rather than bouncing users back out to a traditional website.
Put simply:
Google is wiring AI into Search and Gemini, then reaching out into retailer checkouts.
Microsoft is wiring AI into Edge, Windows and Copilot, then pulling retailers in via merchant programmes and native checkout.
Both are converging on the same destination: agents that discover, compare and buy on behalf of the user.
The new fight for shelf space
For brands, this changes the game in a few important ways.
1. Shelf space moves from “ten blue links” to “one or two agent picks”
When an AI agent does the browsing for you, it does not show ten pages of results. It returns a very short list of options and often a single recommended product.
If Google, Copilot or ChatGPT become the first stop for shopping, the real shelf space shrinks down to:
- the products the agent can clearly understand
- the products the agent is confident enough to recommend
- the products the agent knows it can actually buy, reliably
This is the new AI endcap. If you are not on that short list, you are essentially invisible in that journey.
2. Product data now has to persuade an agent, not just a human
Most brands still write product copy for humans and hope search engines will work it out.
Agentic commerce flips that:
Agents need clean, structured product data: ingredients, sizing, materials, use cases, contraindications, compatibilities.
They need clear benefit language that maps to how people actually ask questions in natural language.
They need consistent pricing and offer data so they can decide whether something is “good value” within a category.
If your PDPs are vague, thin, inconsistent across channels or full of nice brand language but light on specifics, you are giving the agent very little to work with.
3. Operations become part of your ranking signal
Agents are only useful if they can complete the task.
That means they will favour:
merchants with reliable, low friction checkouts
up to date stock and lead time data
clear returns and support policies that can be summarised back to the user
A beautiful product page that frequently leads to out of stock messages or broken carts is no use to an agent that has to deliver a result. Over time, the operational side becomes part of whether you are recommended.
4. Measurement has to catch up
Search teams are still mostly staring at classic SEO dashboards.
In an agentic world, you need to know:
how often your brand appears in AI mode results
whether you show up in agent driven recommendations in Copilot, ChatGPT and others
where you sit versus your category competitors in these new “AI shelves”
That is the data we are obsessed with at LMO7, but the important point is: if you cannot see your AI visibility, you cannot manage it.
Where LMO7 fits into all of this
At LMO7 we are already working with challenger brands who sell on Amazon and D2C to do exactly this:
audit how they show up in AI powered search and shopping
fix the product data and content that agents actually read
line up their marketplaces, D2C and operations with an agent first world
Google’s latest launch is just another signal that agentic commerce is moving out of the lab and into everyday shopping.
If you are a brand and you want to know where you sit on the AI shelf today and what it would take to move up a row, this is the moment to start asking those questions.
What Google has actually shipped
From the outside it looks like a bundle of features. Under the hood it is a single idea: let an AI agent handle more of the journey from “I need something” to “order placed”.
According to TechCrunch and others, the update includes:
Conversational shopping in AI Mode
You describe what you want in natural language and refine it with follow up prompts. Google Search AI Mode turns that into a short list of products, with comparisons and historic pricing data pulled from a Shopping Graph of around 50 billion listings.
Agentic checkout (“Buy for me”)
You pick a product, set the options and a target price. Google tracks the price and, when it hits your number, the agent takes the product to the retailer’s checkout and uses your stored Google Pay details to complete the purchase (with a confirmation step). This starts with partners like Wayfair, Chewy, Quince and selected Shopify merchants.
“Let Google Call” for local stock and deals
You tell Google what you are looking for locally. An AI caller phones around nearby stores, checks availability and promotions, then sends you a neat summary with options.
If you step back, that is discovery, comparison, inventory and checkout all stitched together by an AI layer that is increasingly doing the clicking and dialling on your behalf.
Meanwhile in Redmond: Copilot is not far behind
This is not happening in a vacuum.
Microsoft has already rolled out Copilot Shopping, which helps users discover products, compare prices, track drops and make purchases directly in the Copilot experience.
On the merchant side, Microsoft has launched a Copilot Merchant Program, designed to let brands feed product data into Copilot so their catalogue can surface in AI powered shopping queries. Microsoft
There are also tests of native checkout and Shopify style integrations so that purchases can be completed inside Copilot rather than bouncing users back out to a traditional website.
Put simply:
Google is wiring AI into Search and Gemini, then reaching out into retailer checkouts.
Microsoft is wiring AI into Edge, Windows and Copilot, then pulling retailers in via merchant programmes and native checkout.
Both are converging on the same destination: agents that discover, compare and buy on behalf of the user.
The new fight for shelf space
For brands, this changes the game in a few important ways.
1. Shelf space moves from “ten blue links” to “one or two agent picks”
When an AI agent does the browsing for you, it does not show ten pages of results. It returns a very short list of options and often a single recommended product.
If Google, Copilot or ChatGPT become the first stop for shopping, the real shelf space shrinks down to:
- the products the agent can clearly understand
- the products the agent is confident enough to recommend
- the products the agent knows it can actually buy, reliably
This is the new AI endcap. If you are not on that short list, you are essentially invisible in that journey.
2. Product data now has to persuade an agent, not just a human
Most brands still write product copy for humans and hope search engines will work it out.
Agentic commerce flips that:
Agents need clean, structured product data: ingredients, sizing, materials, use cases, contraindications, compatibilities.
They need clear benefit language that maps to how people actually ask questions in natural language.
They need consistent pricing and offer data so they can decide whether something is “good value” within a category.
If your PDPs are vague, thin, inconsistent across channels or full of nice brand language but light on specifics, you are giving the agent very little to work with.
3. Operations become part of your ranking signal
Agents are only useful if they can complete the task.
That means they will favour:
merchants with reliable, low friction checkouts
up to date stock and lead time data
clear returns and support policies that can be summarised back to the user
A beautiful product page that frequently leads to out of stock messages or broken carts is no use to an agent that has to deliver a result. Over time, the operational side becomes part of whether you are recommended.
4. Measurement has to catch up
Search teams are still mostly staring at classic SEO dashboards.
In an agentic world, you need to know:
how often your brand appears in AI mode results
whether you show up in agent driven recommendations in Copilot, ChatGPT and others
where you sit versus your category competitors in these new “AI shelves”
That is the data we are obsessed with at LMO7, but the important point is: if you cannot see your AI visibility, you cannot manage it.
Where LMO7 fits into all of this
At LMO7 we are already working with challenger brands who sell on Amazon and D2C to do exactly this:
audit how they show up in AI powered search and shopping
fix the product data and content that agents actually read
line up their marketplaces, D2C and operations with an agent first world
Google’s latest launch is just another signal that agentic commerce is moving out of the lab and into everyday shopping.
If you are a brand and you want to know where you sit on the AI shelf today and what it would take to move up a row, this is the moment to start asking those questions.