A Smarter Way to Match Keywords and Search Intent

Keywords are a main building block of Amazon advertising, directly impacting product rankings, traffic, and conversions. Previously, Amazon’s A9 algorithm matched keywords based on product titles, descriptions, and user search terms. However, this approach struggled to handle variations in expression and search intent.

To improve search accuracy and customer experience, Amazon introduced the COSMO algorithm, which focuses on understanding user needs and predicting behavior. However, since Amazon has provided limited public information on COSMO, brands face challenges in adapting their keyword strategies.

How Xnurta AI Helps Brands Adapt to COSMO

To help brands refine their advertising strategies, Xnurta has upgraded its AI-powered keyword recommendation model, introducing two major enhancements:

Fine-tuned AI models for keyword optimization – Uses Transformer-based deep learning to adjust billion-parameter models, ensuring alignment with Amazon’s 300+ product categories across multiple languages.
Integration of Amazon Shopping Queries Data Set – Leverages real Amazon search data to refine keyword accuracy for ASIN-specific attributes like size, model, and color.

These updates enable brands to identify high-potential keywords, adapt to Amazon’s evolving search logic, and improve advertising efficiency.

How Xnurta AI Enhances Keyword Accuracy

1. AI-Powered Keyword Selection Based on Intent

Traditional keyword tools rely on exact phrase matching, making it difficult to capture variations and search intent.

With Xnurta AI, brands can:
Identify synonyms and related terms – Expands keyword targeting beyond exact matches.
Analyze search behavior to detect intent – Matches what customers actually want.

Example: Searching for "Maternity Shoes"

Outcome: Ads reach high-intent shoppers, leading to better CTR and conversion rates.

2. Enhanced Keyword Recognition for Size, Model, and Attributes

Product listings like "10x8x6 FT Outdoor Metal Storage Shed" often contain multiple data points—brand, model, packaging specifications, etc.—but not all of them carry the same weight when it comes to search relevance.

By aligning with Amazon’s COSMO algorithm and fine-tuning our model using Amazon’s eCommerce search dataset, Xnurta significantly improves keyword targeting for products with variations.

Major Upgrade in Size Keyword Variant Recognition

For a product with the size “10x8x6,” the updated model now:

• Accurately recognizes correct variants like "8x10"

• Distinguishes them from incorrect sizes like "8x8" or "8x12"

• Prioritizes correct keywords with higher ranking scores and improved differentiation

This results in more precise keyword recommendations and fewer irrelevant matches.

Broader Reach with Long-Tail and Attribute-Based Keywords

By analyzing the product title, the updated model:

• Surfaces both attribute-related terms (e.g., "outdoor," "metal")

• And hidden user intents (e.g., "small tough shed")

This not only increases targeting accuracy but also expands ad reach by tapping into more long-tail search queries.

Visual Example: Keyword Ranking Comparison Before and After Update

Green = Correct Size
Orange = Incorrect Size  

Expanding Search Coverage with Long-Tail Keywords

Xnurta AI automatically detects and ranks high-performing long-tail keywords, helping brands capture more relevant traffic.

Example: Optimizing a Metal Outdoor Shed

AI-generated keyword recommendations:
"small tough storage shed"
"outdoor storage shed metal" Outcome: Increases ad reach while maintaining relevance.  

Why Xnurta’s AI Keyword Model Matters

FAQ: Xnurta’s AI-Powered Keyword Model

1. How does Xnurta’s AI improve keyword selection?

Xnurta AI analyzes real Amazon search behavior to identify high-converting, intent-driven keywords, ensuring ads reach the right audience.

2. Can Xnurta AI detect product-specific variations?

Yes! It detects product attributes like size, model, and color, ensuring accurate targeting.

Example:
Before AI: "8x8 outdoor shed" (incorrect match for a 10x8 shed).
After AI: "10x8 metal storage shed" (correct match).

3. How often does Xnurta AI update keyword recommendations?

Xnurta AI syncs daily with Amazon Shopping Queries Data, providing real-time keyword optimizations.

Take the guesswork out of keyword targeting.

‍Let Xnurta’s AI adapt to Amazon’s evolving algorithms—so you can reach the right customers with the right keywords, every time.