AI Tagging (or AutoTagging) is a large-language-model approach to completely automatically applying labels to interactions in EdgeTier.
AI Tagging is designed to transform how customer interactions are classified, making tagging more precise and less time-consuming. Built using a Large Language Model (LLM), this feature ensures that every interaction is assigned a tag from a predefined list. This happens automatically and consistently.
Unlike traditional phrase-based tagging or manually selecting disposition reasons, AI Tagging eliminates inconsistencies and guesswork, providing a scalable, accurate, and efficient way to categorise interactions. This approach enhances reporting accuracy, improves operational efficiency, and allows businesses to extract meaningful insights from their customer data with minimal manual effort.
Our initial focus with this feature was solving inaccurate Contact Reason tagging, addressing a key challenge: agents often struggle to select the right category after a customer interaction. Faced with long or complex lists, they may default to generic options, leading to inconsistent data and inaccurate reporting.
To compensate, some customers turned to our Phrase Tagging feature, setting up extensive tag groups and phrase triggers to approximate Contact Reason categorisation. While effective for extracting insights based on nuanced customer language, Phrase Tagging was not designed for “100% coverage” of interactions, requiring extensive setup and maintenance if used for this purpose.
AI Tagging was developed to solve both challenges—providing a structured, automated way to categorise entire interactions without the need for manual selection or complex trigger setups. This results in more reliable reporting, streamlined workflows, and improved data integrity.
Beyond Contact Reasons, AI Tagging can support:
By working closely with you, we define and refine tagging models that align with your specific business needs, ensuring AI Tagging continues to evolve and deliver meaningful insights.
WatchTower has several tagging concepts → System tags are categories that come from the source system that data is read from, Phrase tags are triggered from trigger sentences within interactions, and Auto-Tags are based on an analysis of the entire interaction.
AI Contact Reason Tagging works by analysing conversation transcripts—chats, calls, or emails—to determine the most relevant tag from a predefined set of contact reasons.
Every interaction read into the EdgeTier system is passed through a classification model that maps it to the most appropriate AI tag, using natural language processing (NLP) techniques to assess context, customer intent, and past tagging patterns. If multiple tags are possible for that interaction, a ranking system selects the best match.
We work with you to refine the accuracy of AI tagging through feedback loops, incorporating your corrections and confirmations to improve the system’s future predictions.