Tax Application Virtual Assistant Using LLM Workflow Orchestration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tax applications are complex and require individual training for users, leading to increased time to implementation and decreased productivity due to the need for different levels of assistance.
Innovation Solution
A computing system for progressive virtual assistance in tax applications, which includes a computing device with processing circuitry configured to execute a tax application virtual assistant program. This program uses a chat interface for turn-based dialog sessions, identifies user intents, selects workflows, and utilizes a large language model to generate responses, thereby providing personalized assistance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If individual training is provided for each user to navigate tax applications, then users can efficiently work within the application, but time to implementation increases and productivity in other areas decreases
Solution Approach 1:
A virtual assistant is introduced as an intermediary between the user and the complex tax application system. The virtual assistant handles navigation, explains features, and guides users through tasks, eliminating the need for direct user training on the application's complex interface and workflows.
Solution Approach 2:
The system implements self-service through the virtual assistant, which automatically provides context-aware help and guidance to users based on their current actions and queries. Users receive assistance on-demand without requiring formal training sessions, allowing them to independently navigate the application.
2Ease of operation
If comprehensive assistance is provided to all users, then users can efficiently work within the application, but system complexity and resource requirements increase
Solution Approach 1:
The virtual assistant system is designed to be dynamic, adapting its level of assistance based on user needs, query context, and interaction history. The system adjusts its responses and level of detail provided, offering comprehensive help when needed and concise guidance when sufficient, thereby managing system complexity while maintaining ease of operation.
Solution Approach 2:
The system changes parameters such as the depth of explanation, level of detail in responses, and type of assistance provided based on user characteristics and context. This allows the system to provide appropriate assistance levels without requiring a fully comprehensive system for all possible scenarios.
3Productivity
If individualized assistance levels are provided for each user, then productivity is improved, but the system requires increased adaptability and complexity
Solution Approach 1:
The virtual assistant implements feedback mechanisms where user responses, queries, and interaction patterns are analyzed to dynamically adjust the level and type of assistance provided. This feedback loop enables the system to adapt to individual user needs and improve productivity without requiring manual configuration for each user.
Solution Approach 2:
The system performs preliminary analysis of user queries and context to prepare appropriate levels of assistance before actual interaction. By pre-processing user input and determining the appropriate response strategy, the system can provide individualized assistance efficiently without requiring complex real-time adaptation during user interactions.
Data Source
AI summary
A computing system for progressive virtual assistance in tax applications includes a computing device with processing circuitry configured to implement a tax application virtual assistant program. The processing circuitry is configured to receive a tax-related user query, identify an intent in the query, and select a workflow including operations required for responding to the intent. The processing circuitry is further configured to implement a workflow orchestrator to schedule a sequence of the operations, identify information needed to complete a target operation, and generate and input an information augmentation prompt to a large language model to retrieve the information. The processing circuitry is further configured to receive a response to the information augmentation prompt, generate and input a user query response prompt to the large language model, receive the user query response, and display the user query response in a chat interface.