Next-Gen Sponsored Search Crafting the Perfect Query with Inventory-Aware RAG (InvAwr-RAG) Based GenAI

Jul 16, 2026

Upload by Hazel

Sponsored search drives e-commerce revenue by matching user search queries to advertiser bids on keywords, yet many queries cannot retrieve any sponsored ads due to an ever-changing keyword space, ambiguous intent, and topic variety. To address this missed opportunity, the Inventory-Aware RAG-based Generative AI model (InvAwr-RAG) combines semantic retrieval with real-time inventory data. It generates new and reuses historically successful queries to align with available inventory and campaigns, improving fill rate by 68% while balancing relevance for stronger ad revenue, advertiser ROI, and user experience on Walmart.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

Abstract

Introduction

Sponsored search challenges at Walmart

Related work: IR and NLG-based retrieval

Proposed solution: InvAwr-RAG

Next-Gen Sponsored Search: Crafting the Perfect Query with Inventory-Aware RAG (InvAwr-RAG) -Based GenAI Md Omar Faruk Rokon1, * , Weizhi Du1 , Zhaodong Wang1 and Musen Wen1 1 Walmart AdTech, Sunnyvale, CA, USA Abstract Sponsored search plays a crucial role in e-commerce revenue generation, where advertisers strategically bid on keywords to capture the attention of users through relevant search queries. However, the process of identifying pertinent keywords for a given query presents significant challenges because of a vast and evolving keyword landscape, ambiguous intentions, and topic diversity. This paper highlights an opportunity for to earn a considerable amount of Ads revenue and user engagement where a significant proportion of queries fail to retrieve any sponsored ads. To utilize this opportunity, we introduce the Inventory-Aware RAG-based Generative AI model (InvAwr-RAG), which integrates advanced semantic retrieval and real-time inventory data. This model combines dynamically generated and historically successful queries to align with available inventory and ad campaigns while diversifying rewritten queries to enhance relevance and user engagement. Preliminary results show a significant 68% increase in fill rate and balanced relevance metrics, indicating a strong potential for increased ad revenue. The InvAwr-RAG model sets a new standard in dynamic query optimization, significantly improving ad relevancy, advertiser ROI, and user experience on Walmart’s digital platform. Keywords Dynamic Query Rewriting, Generative AI in Advertising, Sponsored Search, E-commerce Advertising, RAG, 1. Introduction Sponsored search is a cornerstone of revenue generation in e-commerce, where advertisers bid on keywords to display their ads in response to user queries. This system, however, faces significant challenges, including the alignment of user queries with relevant ads—a process complicated by the vast, dynamic keyword landscape and diverse user intents. In the competitive landscape of digital advertising, the efficiency of sponsored search systems is paramount for driving revenue and enhancing user experience on e-commerce platforms like Walmart. A significant challenge that Walmart faces is the presence of search queries that fail to retrieve any sponsored product ads—accounting for approximately 13% of all searches. This issue represents a substantial revenue loss and a missed opportunity to engage potential customers. The inability to show relevant ads not only impacts Walmart’s bottom line but also diminishes the effectiveness of the platform for advertisers seeking visibility and for customers who may miss out on discovering products of interest. Hence, there is a compelling business need for a SIGIR eCom’24: The 2024 SIGIR Workshop On eCommerce, July 18, 2024, Washington, D.C., USA * Corresponding author. $ [mdomarfaruk.rokon@walmart.com](mdomarfaruk.rokon@walmart.com) (M. O. F. Rokon); [weizhi.du@walmart.com](weizhi.du@walmart.com) (W. Du); [zhaodong.wang@walmart.com](zhaodong.wang@walmart.com) (Z. Wang); [musen.wen@walmart.com](musen.wen@walmart.com) (M. Wen) © 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0) . solution that can dynamically align search queries with available inventory and advertising goals, ensuring that every search can result in meaningful ad placements. The core problem this research addresses is the high rate of search queries that yield no ad results due to mismatches between user queries and the current inventory or the specificities of real-time bidding budgets. The challenge is twofold: firstly, to enhance the relevance of ad placements to ensure they correspond with available inventory and meet advertiser bidding strategies; and secondly, to maintain or even improve user experience by presenting ads that are perceived as relevant and potentially interesting. This problem is crucial

FAQs

Why do many sponsored search queries fail to retrieve any ads?

A significant portion of searches cannot match available sponsored product inventory and real-time bidding specifics. The dynamic keyword landscape and diverse user intent further increase the mismatch rate.

What is the InvAwr-RAG model designed to do?

InvAwr-RAG integrates semantic retrieval with real-time inventory data to rewrite incoming user queries so they better match available sponsored products and ad campaigns.

How does InvAwr-RAG generate and improve the query set?

The model combines dynamically generated queries with historically successful queries from search logs. This blending helps align rewritten queries with inventory while improving relevance and user engagement.

Jul 16, 2026

Upload by Hazel