Insight | Nov 20, 2025

This week, we brought together ecommerce leaders to discuss the newly announced Shopify and ChatGPT partnership and what it means for brands preparing for agentic shopping.
Our panelists shared tactical insights on how to get ready for this shift:
- Denise Foley, EVP of Ecommerce at ULE Group
- Purva Gupta, Co-founder and CEO of Lily AI
- Nina Collier, Director of Ecommerce, Third and Grove
Check out the recording: Shopify Meets ChatGPT: Preparing for Agentic Shopping.
Here's what ecommerce leaders need to know right now.
5 ways to get ready to sell on ChatGPT + Shopify
If you're not part of the preview release yet, here's how to prepare:
- Sign up: Submit your interest on both Shopify and ChatGPT's signup pages
- Audit your product data: Ask ChatGPT what feed requirements it needs, then work with your dev team to ensure compliance
- Invest in product content infrastructure: Move away from one-time content creation toward continuous optimization
- Test your feeds: Make sure inventory, pricing, and shipping data is updating in real-time
- Optimize and structure beyond PDPs: Build out educational content, FAQs, and solution guides that help AI agents understand your brand
Your foundation: Intent-led, structured product data
The shift to agentic shopping isn't just about ranking. It's also about reasoning. AI agents need to truly understand your product content to match it with customer intent.
What this means in practice:
Generic product descriptions won't cut it anymore. Instead of "great for all occasions," you need to be specific: is the accessory for a cocktail party, date night, or casual brunch? The machine needs to understand context, not just find keywords.
As Purva explained, your product data needs to be:
- Accurate and unambiguous
- Comprehensive and relevant
- Machine-readable
- Continuously optimized
This isn't a one-time project. Product content needs to be treated as always-on infrastructure, not a static asset you update occasionally.
Real-time data is non-negotiable
ChatGPT expects product feeds to be no more than 15 minutes old. And that's just the starting point. This requirement will only get stricter over time.
During our testing, we saw this play out in real-time. When prompting ChatGPT to find a gift from a specific beauty brand, it immediately asked if we wanted to check product availability. When it couldn't confirm inventory data, it suggested checking Sephora instead.
The takeaway: If your inventory, pricing, and shipping data isn't fresh and accessible, AI agents will route customers to competitors who have their data dialed in.
Your website still matters...maybe more than ever
There's been speculation about whether brand websites will become obsolete in an agentic shopping world. The answer? Your site is more important now than ever.
Your website is the only place where you have complete control over your brand narrative and how your products are presented. On marketplaces, in AI feeds, and through third-party channels, you’re constrained by their formats and rules. But on your own site, you control the story.
While structured feeds send specific data to AI agents, your website remains the ultimate source of truth. It's where agents verify information, build trust, and gather the comprehensive content that goes beyond basic product specs. The more accurate and up-to-date your data is, the more your website gains authority as a source of trusted information for LLMs. Once you gain that trust, LLMs will prioritize your content as a source ongoing – the benefits compound.
This includes:
- Blog posts and educational content
- FAQs and solution guides
- Brand story and values
- Detailed product information beyond what fits in feeds
Both human-readable content (for inspiration and clarity) and machine-readable content (for agents) can and should coexist on your PDPs.
Getting Around Shopify's Metafield Limitations
For brands with technical products or extensive product attributes, Shopify's metafield limits can be challenging. Denise shared a practical workaround: consolidate key attributes into your product descriptions.
Her team exports data from their PIM, identifies which attributes are already in Shopify meta fields, and uses a script to pull remaining attributes into bullet points in the product description. This ensures the data appears in feeds sent to ChatGPT and other channels.
It's a work in progress that requires constant refinement, but it's helping them stay ready as the integration rolls out.
First-party data collection still works
One concern we heard: how do you capture customer data when transactions happen through ChatGPT?
The good news is that orders placed through ChatGPT instant checkout still flow back to Shopify. Unlike Amazon, which keeps customer information locked down, you'll get customer email addresses and purchase data.
The opportunity: The Order Confirmation email will now become the first direct interaction you will have with a customer. Use that touchpoint as a brand-building moment and a chance to opt customers into marketing. Consider setting up dedicated email flows specifically for ChatGPT orders.
What’s next?
Agentic shopping won't replace traditional ecommerce, just like online didn't eliminate in-store shopping (which still accounts for 80% of retail in North America). But it will become a significant channel, and brands that prepare now will have a major advantage.
The work you do to optimize for AI agents also improves your traditional SEO, site search, and overall customer experience. It's an investment that pays dividends across channels.
AI is the worst it will ever be today. These agents will only get smarter, faster, and more demanding. The time to prepare is now.
Want to discuss how your brand can prepare for agentic shopping? Reach out to our team at Third and Grove. We're helping clients navigate this transition and some are already in the ChatGPT preview release.
Drop us a line
Have a project in mind?
Contacting Third and Grove may cause awesomeness. Side effects include a website too good to ignore. Proceed at your own risk.


