Voice Navigation Intent Flow for IVR Usability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Interactive Voice Recognition (IVR) systems are not streamlined for natural human interaction, leading to usability issues due to small keypads, diminished screens, and the need for manual interaction, which can be frustrating and unsafe, especially in activities like driving or biking.
Innovation Solution
The implementation of a streamlined telephone call flow that builds 'intent' into navigation paths, using a combination of speaker-independent and speaker-dependent automatic speech recognition (ASR) to reduce unnecessary delays and improve user experience by allowing more accurate pronunciation and 'opt-out' features, and repurposing 'dead' telephone numbers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual interaction with mobile devices is used (typing, keying, scrolling), then information can be obtained, but safety concerns increase and usability deteriorates during activities like driving or biking
Solution Approach 1:
The patent replaces manual mechanical interaction (typing, keying, scrolling) with voice-based acoustic interaction. Users speak natural language commands instead of manually navigating interfaces, eliminating the need for fine motor skills while maintaining information accuracy through speech recognition technology.
Solution Approach 2:
The system enables users to independently obtain information through voice commands without requiring manual device manipulation. The voice-activated interface allows users to self-serve by simply speaking their requests, freeing their hands and improving safety during activities like driving.
2Stability of the object's composition
If traditional IVR systems with fixed navigation paths are used, then system structure is maintained, but user frustration increases due to unnatural interaction requirements
Solution Approach 1:
The patent transforms the static, fixed navigation paths of traditional IVR into dynamic, adaptive conversation flows. The system uses natural language processing to understand user intent and flexibly navigate through information, allowing the interaction structure to adapt to user needs while maintaining overall system organization.
Solution Approach 2:
The system changes the interaction parameter from rigid menu selection to flexible speech-based intent recognition. By accepting natural language input with varying phrasings and interpretations, the system maintains structural integrity while dramatically improving ease of operation through more natural user interaction.
3Adaptability or versatility
If speaker-independent ASR is used to handle various accents, then system accessibility improves, but recognition accuracy may decrease for pronunciations outside grammar files
Solution Approach 1:
The patent introduces speaker-dependent ASR as an intermediary layer between the user's speech and the core processing system. When speech recognition accuracy is insufficient, the system transitions to speaker-dependent mode that adapts to the specific user's pronunciation patterns, effectively bridging the gap between accent diversity and recognition precision.
Solution Approach 2:
The system performs preliminary speaker adaptation by collecting and analyzing a user's speech patterns during initial interactions. This preliminary action builds a personalized recognition model that improves accuracy for that specific speaker while maintaining the overall speaker-independent architecture's ability to handle various accents.
Data Source
AI summary
Call flow information can build ‘intent’ into call flows having a series of intended steps and one or more alternative steps for selected intended steps. Users can advance in the call flow based on any response other than one or more alternative valid responses for the one intended step that lead in the call flow to one or more alternative steps to the one intended step because every action in the call flow motivates the user to follow the predetermined ordering. A user can proceed from a first intended step in the series of intended steps to a second intended step in the series of intended steps based on a user's response or information known or discovered about the user.