FECT: Factuality Evaluation of Interpretive AI-Generated Claims in Contact Center Conversation Transcripts
This addresses a critical issue for enterprises using AI in contact centers to prevent hallucinations in business decisions, though it is incremental as it builds on existing factuality evaluation approaches.
The paper tackles the problem of evaluating factuality in AI-generated interpretive claims from contact center conversations, where ground-truth labels are often unavailable, by introducing a 3D paradigm and a benchmark dataset (FECT) to improve evaluation methods.
Large language models (LLMs) are known to hallucinate, producing natural language outputs that are not grounded in the input, reference materials, or real-world knowledge. In enterprise applications where AI features support business decisions, such hallucinations can be particularly detrimental. LLMs that analyze and summarize contact center conversations introduce a unique set of challenges for factuality evaluation, because ground-truth labels often do not exist for analytical interpretations about sentiments captured in the conversation and root causes of the business problems. To remedy this, we first introduce a \textbf{3D} -- \textbf{Decompose, Decouple, Detach} -- paradigm in the human annotation guideline and the LLM-judges' prompt to ground the factuality labels in linguistically-informed evaluation criteria. We then introduce \textbf{FECT}, a novel benchmark dataset for \textbf{F}actuality \textbf{E}valuation of Interpretive AI-Generated \textbf{C}laims in Contact Center Conversation \textbf{T}ranscripts, labeled under our 3D paradigm. Lastly, we report our findings from aligning LLM-judges on the 3D paradigm. Overall, our findings contribute a new approach for automatically evaluating the factuality of outputs generated by an AI system for analyzing contact center conversations.