Summary
Conversational AI systems often struggle to balance answering questions, understanding people, and helping businesses without being too rigid. The authors propose a new AI design called GRACE that uses a mix of symbolic reasoning and retrieval-augmented generation to guide conversations toward business goals. GRACE tracks the conversation’s progress carefully to ensure it only updates when visitors give new information and chooses responses that help move the chat forward while keeping visitors happy. They tested this system with real estate and cleaning service chats and found it was very accurate at managing conversation states and handling various scenarios.
goal-oriented conversationretrieval-augmented generationsymbolic AIbusiness goal compilerconversation statemonotonic transitionsvisitor utilityfollow-up question selectionmulti-goal dialogueevaluation metrics
Abstract
Goal-oriented conversational systems must answer factual questions, understand visitor-provided information, and advance business objectives without becoming rigid questionnaires. This paper proposes a Symbolic-RAG-Generative architecture centered on the Goal-oriented Retrieval-Augmented Conversation Engine (GRACE). An instruction-constrained Business Goal Compiler transforms business intent into an immutable objective set, normalized priority vector, canonical questions, and initial state vector. At runtime, GRACE receives the complete conversation history, latest visitor message, current state, and grounded answer generated by a separate RAG component. It updates completion only from visitor-authored evidence and selects one contextually modulated follow-up. The core policy maximizes expected business progress subject to a minimum visitor-utility constraint. We formalize the state, monotonic transitions, source separation, question modulation, and constrained policy; present the reference architecture; and define an evaluation comprising 24 English real-estate and 10 Spanish professional-cleaning conversations, totaling 119 protocol-defined visitor turns. Across both domains, GRACE achieves 84.9% exact state-transition accuracy, 91.6% evidence precision, 89.6% evidence recall, 100% monotonicity, and 94.1% terminal-state accuracy. The evaluation establishes compelling symbolic-state performance across standard, multi-goal, RAG-detour, validation, refusal, and robustness scenarios.