How Agentic Search Is Changing E-commerce Discovery

E-commerce search is shifting fast. In this Treecat Chat conversation, Michelle and a Gen Z guest tested how different marketplaces handle a very simple shopping need: finding a men’s green short-sleeve shirt in size XL for a first job interview. The results revealed something important for resellers: the future of search is moving from keyword matching to context-aware, conversational, agentic search.

Buyer Expectations Are Changing

Michelle started by stepping out of the seller mindset and focusing on the buyer experience. From a buyer’s point of view, the universal needs are simple:

• A good deal
• Lots of selection without overwhelming clutter
• Trust in the platform or seller
• Fast, affordable shipping
• Accurate product info
• Easy checkout
• Reliable customer support

These are basic expectations, but they matter even more as shoppers spend less time scrolling and more time expecting search to “just understand” what they want.

What the Search Tests Revealed

To simulate a real shopping scenario, the team searched across platforms like eBay, Poshmark, Depop, Mercari, Amazon, and ChatGPT. The same request produced very different results.

Some takeaways from the experiment:

• eBay returned mixed relevance, including items that weren’t even green.
• Poshmark showed better color matches, but some listings felt overly polished or possibly AI-enhanced.
• Depop leaned more toward youthful, quirky inventory, but still missed the search intent.
• Mercari and others showed that adding terms like “job interview” can sometimes help, but not consistently.
• Amazon felt more reliable because of its structured seller ecosystem and built-in assistant tools.

The big lesson: traditional marketplace search still struggles to fully understand intent.

What Agentic Search Means for Buyers

Michelle explained that agentic search goes beyond simple keyword lookup. Instead of just matching words, it creates a conversation:

• What size do you need?
• What will you use it for?
• What’s your budget?
• Do you want a brand name?
• How fast do you need shipping?

That kind of search can reduce time spent scrolling and help buyers get closer to what they actually want. Instead of browsing endlessly, shoppers can describe a need in natural language and get more refined, useful results.

ChatGPT was used as an example of this shift: rather than just showing listings, it helps organize search intent and suggest better ways to phrase a query.

What Sellers Should Do Now

For resellers, the message is clear: start preparing for search that understands context, utility, and intent. Michelle shared several practical tips:

1. Think beyond one marketplace
   Search is becoming broader than eBay, Poshmark, Mercari, or Amazon. Listings should be designed to surface across the web.

2. Focus on utility
   Include how the item will be used. For example, “job interview,” “summer office wear,” or “casual weekend outfit.”

3. Optimize for mobile
   Since most shoppers search on phones, the most important information should appear early in the title and description.

4. Use crisp, clean photos
   Mobile shoppers need clear visuals. Simple backgrounds and sharp images still perform best.

5. Include synonyms and related terms
   If you’re selling joggers, consider including terms like sweatpants or lounge pants.

6. Tag by occasion
   Add context like wedding, prom, Father’s Day, office wear, or seasonal use.

7. Write descriptive, story-driven copy
   Listings that explain when and how an item can be used may become more valuable in conversational search.

The Big Takeaway for Treecat Sellers

The future of search is not just about keywords. It’s about how people think, speak, and shop in real life. Sellers who start writing listings for intent, context, and utility will be better positioned as platforms move toward more intelligent search tools.

Michelle’s advice was simple: think like a buyer. If you step outside the seller bubble, you can create listings that are easier to discover now and better prepared for the agentic search future ahead.