Generative Order Agent for Natural-Language Retail Item Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional online concierge systems require users to provide multiple inputs and navigate through multiple interfaces to create orders, increasing interaction time and decreasing the likelihood of including large numbers of items, thus reducing user engagement.
Innovation Solution
An online concierge system employs an order generation agent that processes a user's query, comprising an ingredient identification model, item identification model, and candidate order form creation model, to automatically select items and generate order characteristics, reducing user interaction by leveraging generative models and retailer-specific access controls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users individually identify items through multiple search queries and interfaces, then items can be precisely selected, but user interaction time increases
Solution Approach 1:
The order generation agent performs automated item identification and order creation without requiring users to manually search and select items. The system processes natural language queries and automatically generates orders, making the system serve itself rather than requiring extensive user input for each item selection.
Solution Approach 2:
The patent replaces the mechanical interaction system of multiple search queries and manual selections with an automated agent system that processes natural language. The order generation agent uses generative models to interpret user intentions and automatically identify items, substituting the manual mechanical process of item-by-item selection with automated linguistic processing.
2Adaptability or versatility
If users navigate through multiple ordering interfaces, then comprehensive item selection is enabled, but ease of operation decreases
Solution Approach 1:
The patent merges multiple ordering interfaces and item identification processes into a single unified interface that accepts natural language queries. Instead of requiring users to navigate through separate search results pages and selection interfaces for each item, the system combines all these functions into one conversational interaction point.
Solution Approach 2:
The order generation agent serves multiple functions within a single interface: it processes natural language queries, identifies items, determines quantities, selects retailers, and creates orders. This multi-functional approach eliminates the need for users to switch between different interfaces for different ordering tasks.
3Measurement precision
If multiple search queries are required to identify items, then accurate item identification is achieved, but productivity decreases
Solution Approach 1:
The system performs preliminary processing of the user's natural language query to extract all necessary information about desired items, quantities, and preferences before actually creating the order. The order generation agent proactively identifies items and retrieves relevant information in advance, so that when the order is created, all necessary data is already prepared, eliminating the need for multiple sequential search queries.
Data Source
AI summary
An online concierge system receives a query from a user and leverages a set of models to generate an order based on the query. An ingredient identification model is a generative model that receives the query and generates a set of item categories corresponding to items that are combined to satisfy the query. An item identification model trained on catalogs of items offered by retailers receives the set of item categories as an input and generates a list of items available at a retailer corresponding to the set of item categories. The item identification model may generate multiple lists corresponding to different retailers and select a specific list of items based on list scores determined for each list. A candidate order form creation model generates characteristics of an order for obtaining the list of items that leverages prior orders fulfilled for the user.