If you've noticed changes in how Amazon surfaces products in search results over the past 12 months, you're not imagining it.
Amazon quietly rolled out RUFUS — its AI-powered shopping assistant — and it changed how listings get evaluated, ranked, and recommended to buyers.
The part most sellers don't know: RUFUS reads your images.
Not your keywords. Not your bullet points first. Your images.
This post explains what RUFUS is, how it evaluates your listing visuals, and what you can do today to make sure your images pass the RUFUS test.
Key takeaway
Amazon's RUFUS AI evaluates your listing images for specific visual signals before serving your product to buyers. Missing the wrong image type can cost you ranking — regardless of how good your photography looks.
What is RUFUS?
RUFUS (Retail Understanding Framework for Universal Search) is Amazon's AI shopping assistant, launched publicly in early 2024 and expanded significantly through 2025.
Think of it as a smart layer sitting between buyer search intent and product results. When a buyer searches "comfortable nursing scrubs for long shifts," RUFUS doesn't just match keywords — it evaluates which listings best answer that buyer's implied questions:
- Are these scrubs shown in a real hospital setting?
- Can I see the pockets clearly?
- Is there a size chart so I can order the right fit?
- What makes these more comfortable than standard scrubs?
RUFUS answers these questions partly from your copy — but increasingly from your images.
How RUFUS Reads Your Listing Images
Amazon hasn't published a complete technical specification of how RUFUS evaluates images. But based on what we've observed across hundreds of listings, RUFUS appears to evaluate images against a set of implicit buyer questions — questions that vary by product category.
1. Does the hero image clearly identify the product?
RUFUS needs to understand what the product is from the first image. Pure white or light grey background. Product fills 85%+ of the frame. Sharp focus. No props obscuring the product.
If your hero image is cluttered or the product is small in the frame, RUFUS has lower confidence in what the product actually is — and that affects how it surfaces your listing.
2. Do the supporting images answer use-case questions?
"Can I use this for [specific use case]?" is one of RUFUS's most common question types.
For scrubs: "Can nurses wear these for 12-hour shifts?" — a lifestyle image of a nurse wearing them in a hospital setting answers this visually.
For baby bottles: "Is this suitable for newborns?" — a parent feeding a newborn answers this visually.
Images that show the product in real-world use contexts give RUFUS the evidence it needs to recommend the product for specific buyer queries.
3. Can buyers judge size and fit?
Size anxiety is one of the biggest purchase barriers on Amazon — especially for clothing, accessories, and home goods.
RUFUS actively looks for:
- Size charts with measurement guides
- Scale comparison images (product next to common objects)
- Dimension labels in infographic images
- "Fits [specific product/size]" callout images
Without these, RUFUS cannot confidently answer size-related buyer questions — and will deprioritise your listing for those queries.
4. Are product features clearly communicated visually?
Feature callout images — infographics with labels, arrows, and text pointing to specific product features — help RUFUS connect your product's features to buyer intent.
"Does this have pockets?" → answered by an infographic showing pocket locations and sizes.
"Is the material breathable?" → answered by a close-up with "4-way stretch fabric" label.
"Is this BPA-free?" → answered by a callout image showing the certification.
5. Are trust signals visible?
RUFUS evaluates buyer confidence signals. This includes certifications visible in images (FDA, BPA-free, CE, etc.), guarantee badges, and awards or press mentions shown visually.
Important
Mentioning certifications only in your bullet points is not enough. RUFUS evaluates visual trust signals separately. If your certifications aren't visible in your images, RUFUS scores your listing lower on trust — even if the bullets mention them.
What is COSMO — and Why It Matters Too
Alongside RUFUS, Amazon uses a semantic AI model called COSMO that evaluates how well your listing covers the semantic relationships buyers use to discover and evaluate products:
- Use case coverage: Is it clear when and where this product is used?
- Target audience clarity: Is it clear who this is for?
- Problem/solution framing: Does the listing state the problem it solves?
- Occasion context: Does it mention events, seasons, or occasions?
- Complementary products: Does it reference what pairs with it?
COSMO analysis applies primarily to your copy — but image alt text, infographic text, and A+ content text all feed into the semantic picture COSMO builds of your listing.
See your RUFUS score in 30 seconds
Enter any Amazon ASIN and get your free Visual Authority Score — including RUFUS gap analysis specific to your product.
Get My Free Score →The Visual Authority Score
To make RUFUS and COSMO analysis practical for sellers, we built the Visual Authority Score — a 0–100 score that quantifies how well your listing images perform against these AI signals.
The VAS scores 10 visual dimensions:
| Dimension | What it measures |
|---|---|
| Hero Image | White bg, fills frame, product clarity and professionalism |
| Infographic | Feature callouts with labels, arrows, mobile-readable text |
| Lifestyle Shot | Real people in real-world use contexts |
| Benefit Callouts | Value proposition shown visually, not just in copy |
| Size Reference | Dimension charts, scale comparisons, measurement labels |
| Trust Signals | Certifications, badges, guarantees shown visually |
| Mobile Optimised | Text and details readable at thumbnail size |
| Branding | Consistent colours, logo, professional design identity |
| Video Ready | Video content present in listing |
| A+ Content | Enhanced brand content detected |
A score below 50 is Critical — significant visual gaps hurting your RUFUS performance. 50–70 is Developing. 70–85 is Competitive. 85+ is Authority.
How to Optimise Your Listing Images for RUFUS
Based on our analysis of hundreds of listings, here are the highest-impact changes you can make:
Priority 1: Add a real lifestyle image
If you have zero lifestyle images, this is the single highest-impact change. RUFUS needs to see your product in use to answer use-case queries. A product on a white background tells RUFUS nothing about who uses it or when.
Good lifestyle image: Real person (not stock) using the product in the setting your buyers are in.
Priority 2: Add a size reference image
Size anxiety kills conversions. One clear size chart or scale comparison image can significantly reduce return rates and improve buyer confidence. For clothing: bust/waist/length measurements. For accessories: product next to a hand for scale.
Priority 3: Add infographic callouts
Text callouts on images let RUFUS read your product features visually. Good infographic: 4–5 feature callouts pointing to specific product features with clear, mobile-readable text. Minimum 12pt equivalent font size.
Priority 4: Add a trust signal image
Even one image showing your product's key certifications, guarantees, or quality standards builds RUFUS confidence. BPA-free, FDA approval, satisfaction guarantees — show these visually, not just in bullets.
Priority 5: Improve mobile readability
Over 60% of Amazon purchases happen on mobile. Text in your infographic images must be readable at thumbnail size. If buyers have to zoom in to read your callouts, RUFUS scores your mobile optimisation low.
Your RUFUS image optimisation checklist
How to Check Your RUFUS Performance
The easiest way to see exactly which visual gaps are hurting your RUFUS performance is to run a free scan at app.letnexus.com.
Enter your ASIN — the AI downloads your images, scores each dimension with evidence, and shows you exactly which visual gaps RUFUS has identified, which buyer questions your images can't answer, and which image types to create first for maximum impact.
The scan takes 30 seconds and requires no account.
Summary
RUFUS is reading your listing images — whether you've optimised for it or not. The sellers who win on Amazon's AI-powered search in 2026 are the ones who understand what RUFUS looks for and create images that answer the buyer questions RUFUS is scoring.