Hybrid AI-Human Virtual Assistant Architecture for Customer Service

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current virtual assistant systems for customer service face challenges in understanding consumer interactions, leading to frustration and increased costs due to the need for extensive re-prompting and human intervention, which affects consumer loyalty and Net Promoter Scores.

Innovation Solution

A system that combines Artificial Intelligence (AI) and Human Intelligence (HI) to provide seamless customer service by using AI to automate tasks while leveraging human agents for complex interactions, dynamically adjusting based on consumer needs and business requirements, and utilizing machine learning to improve AI capabilities over time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI systems are used to automate customer service conversations, then productivity increases and costs decrease, but understanding accuracy deteriorates leading to frustration and increased human intervention

Engineering Contradiction:

Improveautomation capabilityVSAvoidunderstanding accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent introduces a hybrid architecture where human agents serve as intermediaries between the AI system and customers. When the AI's understanding accuracy falls below a threshold, the conversation is seamlessly transferred to a human agent who can provide reliable understanding. This mediator approach maintains high productivity for routine tasks while ensuring reliability when needed.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system dynamically adjusts the level of automation based on conversation context and AI confidence levels. Rather than a static automated-or-human approach, the system fluidly transitions between AI handling and human intervention based on real-time performance metrics, optimizing both productivity and reliability adaptively.

Inventive Principle:

Principle #15Dynamics

2Reliability

If extensive re-prompting and confirmation are used to verify AI understanding, then understanding accuracy improves, but loss of time increases and customer experience deteriorates

Engineering Contradiction:

Improveunderstanding verificationVSAvoidconversation duration

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary verification by continuously monitoring AI confidence levels during the conversation. Instead of waiting for customer frustration or using extensive re-prompting, the system proactively detects when understanding is insufficient and prepares for seamless human handoff, reducing unnecessary time loss.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements continuous feedback loops where AI performance is monitored in real-time. When understanding accuracy falls below thresholds, the feedback triggers automatic escalation to human agents. This feedback mechanism verifies understanding efficiently without requiring lengthy re-prompting sequences.

Inventive Principle:

Principle #23Feedback

3Productivity

If fully automated systems are deployed for simple tasks, then productivity increases, but adaptability decreases when handling complex or unexpected consumer inquiries

Engineering Contradiction:

Improveautomation efficiencyVSAvoidhandling complexity

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent segments customer service tasks into automated-handleable portions and human-handleable portions. Simple, routine tasks are segmented for AI automation to maximize productivity, while complex or unexpected inquiries are segmented for human agent handling to maintain adaptability. This segmentation allows the system to optimize for both efficiency and versatility.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The hybrid system serves multiple functions: it acts as a fully automated system for routine tasks, a monitoring system for quality control, and a seamless handoff mechanism for complex cases. This multi-functionality allows the same infrastructure to handle both simple automation and complex adaptability requirements.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Reliability

If human agents are used for all complex interactions, then understanding accuracy improves, but productivity decreases and costs increase

Engineering Contradiction:

Improveconversation qualityVSAvoidservice capacity

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

Instead of deploying human agents for all interactions (excessive action), the system uses human agents only partially—specifically for complex or AI-failed cases. This partial deployment maintains high conversation quality when needed while preserving overall productivity by keeping most routine interactions automated.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12003668B1Virtual assistant architecture for natural language understanding in a customer service system

Publication Date: 2024.06.04 INTERACTIONS LLC (US)

AI summary

A virtual assistant system for communicating with customers uses human intelligence to correct any errors in the system AI, while collecting data for machine learning and future improvements for more automation. The system may use a modular design, with separate components for carrying out different system functions and sub-functions, and with frameworks for selecting the component best able to respond to a given customer conversation.