Ameritech's call centers collected large volumns of verbatim comment data describing customer problems. This data contained valuable business intelligence, which could not be exploited because of the vast amount of data.

Typical databases contained 1 million commented records and grew by thousands of records each day. Due to the enourmous volume of data, manual reading was impractical and statndard statistical and data mining tools are not applicable to textual data.

A text mining tool was designed to automatically analyze comment data to extract most significant trends. Statistical NLP (Natural Language Processing) techniques were applied to automatically extract and summarize the most salient concepts in the data. This is a technique based on AI text processing research and is also used to analyze text responses to surveys.