Look, something insane is happening on Walmart and not on Amazon.

Millions of shoppers are now asking Sparky to do all the heavy lifting. The days of casual browsing listing after listing are coming to a close.

We are entering the era of “Hey Sparky, find me the best wireless black keyboard with a 4.5-star rating or higher.” Now, the question is this: will you adapt or whine about it?

Rich media has lived on Walmart for years, but most brand owners have done it WRONG. What changed is the WEIGHT it carries. Every module you add gives AI more concrete signals about what the product is, who it serves, and why it solves a problem. Skip it, and you’re feeding the assistants thinner, weaker material than your competitors. Avoid this like the plague. Trust me.

What “Rich Media for AI” Actually Means

It is structured content designed so both on-site Sparky and external AI models can parse it cleanly. Videos, 360 spins, FAQs, comparison tables, feature sets, and documents all count. Each one expands the dataset AI uses to decide whether to recommend your item to a Walmart customer.

It shows up in two zones:

• Above the fold: Videos and 360 spins sit at the top, visible the second the page loads.
• Below the fold: The brand highlights section holds the deeper modules that tell the full product and brand story.

Both Sparky and outside models scan the whole page. Outside the basics (title, description, key features), rich media is one of the highest-leverage ways to hand AI more usable context.

Why Richer Content Wins for Sparky and Beyond

AI optimization breaks into three practical pillars:

• Nail the fundamentals. Strong sales history and clean data still matter most because algorithms evolve gradually.
• Optimize what Sparky sees on-site. It reads the full listing plus your sales and purchase history.
• Optimize for off-site AI. ChatGPT, Gemini, and Google AI Overviews may only grab pieces of the page—titles, images, or visible modules.

Rich media feeds both channels. Ignoring it costs you in three direct ways:

• Weaker rankings. Higher conversion from better content lifts search performance.
• More returns. Shoppers who understand the product before buying send fewer items back.
• Less dwell time. Rich modules keep people on your page longer before competing ads and products appear.

Walmart also favors sellers who bring external traffic onto the platform. Helllloooooo TikTok influencers! Off-site AI recommendations that land on your listing help Walmart win more of the broader e-commerce fight.

What Actually Shifted in 2026

Search moved from a single query to a conversation. Someone used to type “sunscreen.” Now they describe their skin type, past products that failed, and specific concerns, and AI picks a product. Once they land on a listing, they keep asking: Does it have this ingredient? Is it kid-safe? What do the reviews actually say?

Two other changes matter for planning:

• Personalized results. Search and homepage placements now surface different products based on buying history. Last holiday season, sales spread across more SKUs because shoppers saw matches for their own categories.
• Off-site influence. Many shoppers arrive already leaning toward a product after a ChatGPT session, Google AI Overview, or TikTok/YouTube video. That path from AI conversation straight to a Walmart listing keeps growing.

Above-the-Fold vs. Below-the-Fold: What Each Hands AI

Above-the-fold modules (video and 360) give immediate visual proof and fast signals. Below-the-fold modules in the brand highlights section deliver deeper text and structured information. Both zones get read. The below-the-fold space is where most of the high-impact modules live, and most sellers still leave it empty.

How to Actually Optimize with Rich Media

Prioritize the modules that move AI the most, then layer in category-specific ones.

1. Lead with FAQs
   This is the highest-impact module by a wide margin. AI operates through questions and answers; an FAQ module puts that exact format on the page. Strong answers cover what the product is for, who it is best for, and the specific problem it solves. Write in natural phrases (“nutritious bar for kids who won’t sit still”) instead of bare keywords. Keep it human-readable. AI scores for that. Brand colors and formatting are fine as long as the text stays clear.
2. Use comparison tables to tell the catalog story
   Place your product next to a few others from your own line. This shows AI (and shoppers) who you serve and how the items differ. It works especially well when you have a proven seller and a new launch—let the strong item pull traffic toward the new one. It also keeps more of the session on your brand instead of leaking to competitors.
3. Build feature sets and image carousels for depth
   Feature sets combine images, text, and video into the most complete explanation of the product or brand. Image carousels can reuse existing shots or add new ones. Both give AI additional media beyond pure text. Rich, detailed content signals that the brand invested in the listing. The kind of signal AI prefers when recommending products.
4. Add subtitled video and a 360 spin
   You can host multiple videos; put the strongest ones first. Some AI models scrape the video itself; many do not. The video title and the subtitle file you upload matter either way—they add more readable text for AI. Shoppers watch the video too. A 360 spin lets people (and AI interpreting the page) see the full product quickly without extra clicks. For longer videos, a below-the-fold video carousel gives them space.
5. Upload product documents
   Host instruction manuals, spec sheets, or brand info as PDFs right on the listing. Shoppers get answers without hunting for paper inserts. AI gets a full, structured source of knowledge so it can answer more detailed follow-up questions accurately.
6. Match the module to the product
   Not every format fits every item.
   • Before-and-after for cleaning products or fitness gear.
   • Product tours with hotspots for sets or multi-ingredient items.
   • Brand story or mood board to show who you are and who you serve.
   • Extra 360 viewers when a closer look below the fold helps.

Choose based on what the shopper actually needs to understand before buying.

Mistakes That Quietly Kill AI Performance

• Skipping fundamentals. Rich media amplifies a strong listing; it does not fix a weak one.
• Conflicting information. Title, description, and features do not need to repeat each other, but they cannot contradict. When they do, AI cannot confidently identify the product and moves on.
• Filling only the required attributes. Walmart pushes listing data to other platforms. More complete attributes create more chances to surface.
• Writing for robots instead of people. Keyword-stuffed copy reads poorly and gets scored lower by AI that prioritizes human-readable content.
• Describing the product but not the shopper. AI matches products to people and situations. Every module should make clear who it is for and in what context.

Bottom Line

AI does not stop at your title. It reads the whole page. Every solid module you add gives Sparky and outside assistants more reasons to understand the product and recommend it. Start with FAQs, add a comparison table, then layer in feature sets, subtitled video, documents, and the modules that fit the specific product. Treat rich media as the natural extension of clean fundamentals, not a replacement for them. The sellers who fill the page with usable, human-readable content are the ones AI will keep surfacing.

It’s really that simple!

Speaking of AI...

I know I covered some of this a few weeks ago, but it honestly deserves more attention.

As you already know, Amazon has blocked AI shopping agents, but Walmart has opened up the floodgates.

On October 21, 2026, it announced the Agent Gateway, Agent Profiles, and new rules for how external AI agents can shop on its platform.

Here’s the simple version:

External AI agents (like ones from Meta, OpenAI, or others) no longer need to browse Walmart’s website like a human. Instead:

• They connect through the Agent Gateway (built on Anthropic’s Model Context Protocol / MCP).
• Once inside, they talk to Sparky — Walmart’s own shopping agent.
• The customer gets an Agent Profile inside their Walmart account. This gives the agent its own identity, order history, and spending rules.

Think of it like adding an authorized user to a credit card. The customer can require approval for every purchase or set a spending limit. Agents never see the customer’s card details.

Why This Matters

Until now, retailers mostly had two bad options: let an agent log in with the customer’s credentials (risky), or block it entirely. Walmart created a middle path. Agents get controlled access, customers stay in charge, and Walmart keeps the data inside its own walls.

Importantly, visiting agents (and the companies behind them) can only use what they see for the specific shopping task. They can’t take the data to train models, target ads, or build their own analytics products.

The Protocol Behind It

Walmart helped launch something called the Personal Agent Protocol (nicknamed Poppy or PAP) earlier in October with Meta, Sierra, and others. It’s an open standard for how a personal AI agent should identify itself, sign in on the retailer’s site, and get only the permissions the user allows. Walmart’s Gateway doesn’t officially say it uses PAP yet, but the ideas line up closely.

Ads and Recommendations

Sparky already has ads. When an outside agent shops through Sparky instead of a normal results page, any sponsored recommendations will have to appear inside the answer Sparky gives the agent. Walmart hasn’t confirmed how this will work yet, but the logic is clear: if no human is looking at a page, the ad has to live in the conversation.

Shoppers seem relatively open to this if it’s labeled. A Koddi survey found most people are fine with brands paying for a spot on an AI shortlist as long as it’s transparent.

The Bigger Picture

Amazon is currently blocking agents it doesn’t control. Walmart is building them a controlled entrance and routing everything through Sparky. The data, the “usuals” (repeat purchases), and the recommendation layer stay with Walmart. And guess what helps Sparky the most? Everything I mentioned above in the hack of the week…

For brands, this means the first purchase and strong, clear product data matter even more. Agents will look for reliable matches and in-stock items. Being someone’s “usual” becomes harder to break into if the memory lives inside Walmart’s system.

Bottom line: Walmart isn’t just allowing AI agents.

It’s forcing them to talk to Sparky under its rules, while keeping customer data and the shopping relationship under its own control. Amazon is still saying no. Walmart is saying yes, but on its terms.

I give it 12 months before Amazon gives up and allows agents to start shopping. or maybe I’m being too conservative? Amazon can’t ice them out forever. Walmart was very smart to let them in early.