Automated Agents With Clickable Prompts for Online Browsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing large language model based chatbots require users to type long prompts and specify product references, leading to time-consuming interactions that can frustrate users and reduce conversion rates on e-commerce websites.
Innovation Solution
Implementing clickable prompt buttons that interact with large language models, allowing users to quickly receive product information by clicking on pre-programmed buttons that provide contextual information, eliminating the need for manual input and long searches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If large language model based chatbots are used to provide product information, then users can access large amounts of information, but users must type long prompts and specify product references which is time consuming
Solution Approach 1:
The system pre-generates and displays multiple prompt options related to the product before the user needs to ask. These prompts are created based on the product context and are ready for immediate selection, eliminating the need for users to type out their questions from scratch.
Solution Approach 2:
The patent introduces an intermediary layer between the user and the large language model. Instead of direct user-to-LLM interaction requiring typed prompts, the system provides a middle layer of pre-generated prompt options that mediate the interaction, making it faster and more convenient while still leveraging the LLM's capabilities.
2Ease of operation
If traditional browsing interfaces are used, then users can navigate through pages, but users must click through endless pages and search through long product detail pages which is frustrating
Solution Approach 1:
The system segments the information retrieval process by providing discrete, clickable prompt options instead of requiring users to navigate through continuous pages. Each prompt represents a segmented, focused question about the product, allowing users to quickly jump to specific information categories without scrolling through long pages.
Solution Approach 2:
The patent adds a new dimension to the browsing interface by overlaying interactive prompt buttons on the product page. This creates a vertical layer of interaction above the traditional horizontal page navigation, allowing users to access information through a different dimensional approach - clicking prompts rather than scrolling pages.
3Productivity
If clickable prompt buttons are implemented to simplify user interaction, then users can quickly access information by clicking, but the system must process and generate multiple prompt types which increases system complexity
Solution Approach 1:
The system performs preliminary processing by pre-generating multiple prompt options based on the product context before the user interacts. This upfront work, though complex, is done once per product context and then stored as ready-to-use prompt options, reducing the complexity of real-time processing when users click.
Solution Approach 2:
The system manages complexity by changing parameters of the prompt generation process - using different prompt templates and parameters for different types of information requests. By organizing prompts into types with different parameters, the system can efficiently handle various information needs without creating entirely separate processing paths for each scenario.
Data Source
AI summary
A system and associated method for simplifying and enhancing a customer browsing experience on websites. In an embodiment, the method involves, upon determining that a customer has clicked on a clickable prompt button embedded at a website, in association with a specific product or article, displaying instantaneous product information to the user without requiring the user to type in lengthy requests. In another embodiment, upon determining that a customer has clicked on a clickable prompt button embedded at a website, providing means for enabling the customer to interact with a large language model that is capable of responding with specificity to user questions about products on the websites based on contextual information transparently provided by the clickable prompt buttons.