Why agentic shopping matters
AI-driven traffic to Shopify stores has grown 7x in the past year, and 64% of shoppers are likely to use AI when making a purchase. As AI increasingly shapes how customers discover products, brands need to ensure they show up clearly and accurately in these experiences.
Agentic shopping includes any type of discovery, browsing, or purchasing of goods or services done through AI. Agentic shopping can be done in a large language model (LLM) like ChatGPT, Gemini, or Claude, or through AI agents that are on your brand website.
No matter where agentic shopping happens, optimizing for it starts on your website. A clear, well-structured site builds shopper confidence and provides the context AI needs to accurately recommend your products.
Understand agentic shopping
So, what is agentic shopping and how does it impact the shopping journey? Watch the following video to find out.
Optimize your website for agentic shopping
Now that you know what agentic shopping is, it’s important to optimize your website so that your brand and products can be AI discoverable.
Industry expert, Katherine Boyarsky, recently shared best practices on how to optimize website content for agentic shopping. Let's run through a few of these tips in the gallery below. (Source: Klaviyo Blog)
Test how your brand appears in AI platforms
Put yourself in your customer’s shoes and ask LLMs the questions they might be asking. This allows you to see how AI currently understands your products and where important information may be missing or unclear.
What to do:
Run conversational, shopper-style searches in AI platforms to see how your products currently appear. Turn off your AI’s memory, or search in incognito mode so that your results reflect what a first-time shopper would see. Document what AI gets wrong or misses from your brand or products.
Guiding questions:
- What problems are your customers trying to solve?
- What language do your customers use?
- What matters most to your customers? Price, comfort, convenience, etc.?
Example:
What’s a lightweight, waterproof jacket under $100?
How to:
Review your visibility monthly to ensure your brand continues to appear. Similarly, if your brand doesn’t appear now, make the changes recommended in the following steps and review your visibility afterwards.
Audit your website's product data
Structured product data acts like a label for your product that AI and search platforms read to display accurate details. Most brands already do a great job covering the basics.
Some examples of general product data include:
- Product name
- Product category
- Short product description
- Price
- SKU or product ID
- Product images
- Variant details (such as size, color, or material)
What to do:
Build on your website’s existing foundation by including advanced product data to give AI richer context about your products.
Advanced product data:
- Care instructions (dishwasher safe, store in a cool environment, etc.)
- Compatibility (works with…)
- Use cases
- Safety recommendations (age range, allergen information)
Example:
Product: waterproof jacket
- Technology: GORE-TEX 2-Layer waterproof membrane
- Best for: day hikes, backpacking, alpine travel, rugged outdoor use
- Shell material: abrasion-resistant nylon designed to handle brush and rock
- Weight: lightweight shell (14 oz. - size medium)
- Care instructions: machine wash cold, air dry
How to:
Add advanced product data (metafields) to strengthen your product information. This can be on the product description page (PDP) or on another page on your website.
Use an AI platform to compare your product to competitors. Note what information is missing from your product and add features or attributes into your product page.
Update your website product catalog details
Your product catalog should reflect how shoppers search and ask questions.
What to do: write product content in natural, conversational language that reflects how real shoppers search for products and ask questions. Highlight specific use cases, benefits, and customer pain points.
Example: “Looking for a lightweight waterproof jacket that can handle real rain without weighing you down?” This GORE-TEX rain jacket is built for hikers and backpackers who need reliable protection in wet, windy conditions. It’s fully waterproof and breathable, so you stay dry on long hikes without overheating.
Ideal for day hikes, backpacking trips, and unpredictable weather, the lightweight shell packs easily into your bag and layers comfortably over a fleece or insulated jacket. Perfect if you want durable rain protection that won’t feel bulky or restrictive on the trail.”
How to:
Use AI tools to generate questions that your website should answer. Then, update your product descriptions or create additional information pages that help answer these questions.
Publish AI-optimized educational content on your website
AI tools often pull from FAQs, guides, and educational content when recommending products.
What to do: create educational content on your website that answers real customer questions.
Example:
Q: Is this jacket waterproof enough for long hikes in heavy rain?
A: Yes. This jacket is built for extended exposure to rain during activities like hiking and backpacking. It uses a GORE-TEX® waterproof membrane with fully taped seams to keep water out, even in steady or heavy rain. It’s also breathable, so moisture can escape while you’re moving, helping you stay dry and comfortable on longer hikes.
How to:
Identify common questions that your customers ask. Look at DMs, emails, reviews, or support tickets to reveal trends or common themes.
Encourage reviews and user generated content (UGC)
Customer reviews help shoppers evaluate products, and give AI real-world signals about the product.
What to do: encourage customers to leave detailed, descriptive reviews that explain who they are, what they needed, and how the product performed.
Example:
5 stars
“I bought this jacket for a weekend backpacking trip in the Pacific Northwest, and it held up incredibly well in steady rain and wind. I needed something lightweight that wouldn’t take up much space in my pack, but still tough enough for potential rocky terrain and hiking.
The GORE-TEX shell kept me completely dry during a six-hour hike, and it never felt clammy even when I was moving uphill. I layered it over a fleece and had no issues with fit or mobility.”
How to:
- Set up an automated review request flow
- Offer incentives for detailed feedback that includes photos and/or videos
- Guide customers with tailored questions to make reviews more helpful
For off-site UGC:
- Periodically search your brand or category on YouTube and Reddit
- Look for comparison videos, review explainers, or discussion threads
- Use these insights to identify common questions, objections, or language customers use
Next steps:
After optimizing your website for agentic shopping, the next step is to understand how shoppers arrive from AI tools.
When someone clicks a link to your site from an AI platform like ChatGPT, Shopify captures the referral source as utm_source=chatgpt.com.
Tracking this traffic helps you understand where shoppers are entering the journey, and what they do next, so that you can follow up with relevant, timely messages.
What to do:
Track traffic coming from LLMs to see where your customers are coming from, and learn where they are in the shopping journey. Then, use this information to create follow up flows to bring potential shoppers through the customer journey.
How to:
- Identify AI-driven traffic in your analytics. In Shopify or your analytics tool, filter traffic by referral or UTM source (utm_source=chatgpt.com) to see customers coming from AI platforms.
- Review on-site behavior from visitors. Look at key actions like product views, and add-to-cart to see where AI-driven shoppers are engaging, or dropping off.
- Segment AI-driven visitors in your marketing platform. Create a segment of visitors who arrived from LLMs so that you can treat them differently from other traffic sources.
- Build follow up flows. Use behavior signals to trigger relevant flows such as browse abandonment, cart abandonment, post-purchase.