Next Best Action Recommendation Engine Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional enterprise applications lack the ability to understand user behavior in real-time and support responsive user navigation, leading to inefficiencies in user interface performance and user experience due to complex computational models.
Innovation Solution
The implementation of a system that uses a machine learning model to determine Next Best Actions (NBA) based on user interaction statistics and context, providing personalized recommendations to users through a responsive user interface, while leveraging a UI generation engine and data repository to analyze and present these actions effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a complex computational model is used to support user interface components, then the accuracy of user behavior analysis is improved, but the page rendering time increases
Solution Approach 1:
The system pre-generates multiple UI variations and pre-calculates their associated metrics (engagement, conversion, time-on-page) before actual user interaction occurs. This allows the complex computational model to be executed in advance, storing results that can be quickly retrieved during page rendering without impacting real-time performance.
Solution Approach 2:
The computational model is divided into separate processing stages: data collection, analysis, prediction generation, and UI rendering. Each stage operates independently with defined inputs and outputs, allowing the complex analysis to be performed asynchronously while the UI rendering proceeds with pre-computed results, eliminating the bottleneck effect.
2Ease of operation
If personalized recommendations are provided to each user, then user experience is improved, but the computational complexity increases
Solution Approach 1:
The system adjusts the level of personalization dynamically based on user profile attributes, interaction history depth, and context relevance. For new users with minimal data, generic recommendations are provided; for established users with rich interaction patterns, highly personalized recommendations are generated, optimizing the computational resources required.
Solution Approach 2:
Instead of performing full computational analysis for each user query, the system creates simplified copies or proxies of user profiles with pre-extracted features and preferences. These proxy profiles enable fast recommendation generation without repeatedly executing complex analysis algorithms on the full user data set.
3Speed
If real-time user interaction analysis is performed, then the responsiveness of the interface is improved, but the processing load increases
Solution Approach 1:
The system implements periodic sampling of user interactions rather than continuous analysis. Key interaction events (clicks, hovers, form submissions) trigger analysis cycles, while passive scrolling or minor interactions are batched or skipped. This periodic approach maintains responsiveness for significant events while reducing overall processing load during low-activity periods.
Data Source
AI summary
Techniques for determining a Next Best Action (NBA) are disclosed, with the determination being based on a position within an application, past actions by a user when experiencing a similar application context, and/or tasks in queue for the user to execute from current state of the application. Techniques are also disclosed for displaying an interface that includes the NBA in conjunction with a specific rationale for presenting the NBA, such as recommendations by a trusted person, based on the user's and/or other users' past behavior, and descriptions developed based on the specific NBA. Also, methods for determining a best NBA are disclosed, with the NBA being selected by applying static rules to a data set, heuristically analyzing the data set, and/or applying a machine learning model to the data set.