Generative AI Recommendation Engine for Personalized Item Mapping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional item listing systems lack the comprehensive logic and infrastructure to effectively provide generative-AI-based recommendations, leading to generic and non-personalized suggestions that fail to align with individual user preferences and trends.

Innovation Solution

The implementation of a generative AI recommendation engine that utilizes a large language model to generate review-based recommendation guides, incorporating user feedback and reviews to provide personalized recommendations and map user preferences to relevant items.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If conventional recommendation systems are used in item listing systems, then the system structure remains simple, but the recommendations become generic and fail to align with individual user preferences

Engineering Contradiction:
Improvepersonalization of recommendationsVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a generative AI model as an intermediary component between the item listing system and user preferences. This model processes review data and generates personalized recommendation guides, enabling adaptive recommendations without requiring complete system restructuring. The intermediary handles the complexity of personalization internally while presenting simplified personalized results to users.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent extracts the recommendation generation function into a separate generative AI model that can be independently trained and updated using review data. This extracted component processes user feedback and preference information separately from the core item listing system, allowing the main system to remain relatively simple while the extracted recommendation engine provides sophisticated personalization.

Inventive Principle:
Principle #2Taking out (Extraction)

2Manufacturing precision

If review data is extensively processed to generate personalized recommendations, then recommendation quality improves, but processing time and computational resources increase

Engineering Contradiction:
Improverecommendation qualityVSAvoidprocessing time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent implements preliminary processing of review data to pre-generate recommendation guides based on aggregated user feedback. By performing this processing in advance rather than in real-time during user interactions, the system builds a repository of pre-computed recommendations that can be quickly retrieved and personalized, significantly reducing processing time while maintaining high recommendation quality.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent employs dynamic processing strategies where the depth and extent of review data processing adapt based on user context, device capabilities, and time constraints. The system can adjust between more comprehensive processing for quality-critical scenarios and faster processing for time-sensitive situations, optimizing the balance between recommendation quality and processing time.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4571621A1Generative artificial intelligence recommendation engine in an item listing system
Publication Date: 2025.06.18 EBAY INC
  • EP4571621A1 patent drawingFigure 1A
  • EP4571621A1 patent drawingFigure 1B
  • EP4571621A1 patent drawingFigure 1C

AI summary

Methods, systems, and computer storage media for providing generative artificial intelligence (AI) recommendation management using an artificial intelligence system in an item listing system. A generative AI recommendation engine supports generative AI recommendation management based on a review-based recommendation platform including offline generative AI operations, review-based recommendation guides for items and a review-based recommendation logic. Using generative AI techniques, review-based recommendation guides are generated for a plurality of items. The review-based recommendation logic supports identifying review-based recommended items for users based on review data and the review-based recommendation guides. In operation, review data - associated with a user - for a first item, is accessed. Based on the review data, a review-based recommendation guide feature of the first item is identified. The review-based recommendation guide feature of the first item is mapped to a review-based recommendation guide feature of a second item. The second item is communicated as a review-based recommended item.