System, method, and computer program product for automatically analyzing and categorizing phone calls
Disclosed embodiments relate generally to systems and methods for automatically analyzing and categorizing phone calls for subsequent processing by telecommunications systems. Waveform and/or data analysis may automatically be performed on incoming call records in real or near real time.
1. A system for automatically processing call records comprising:
a call records server that provides call recordings in real time upon completion of the call for analysis,
a call records analyzer that provides parallel processing of waveform and transcription of the call records;
means for categorizing the calls by processing structured features and unstructured data features through vectorization of the transcription so that the call records are objectively tagged and stored in groups of predetermined call categories, and
a display, wherein selection of a stored call record displays specific data for the call including a visual waveform depicting a breakdown of the call participants, a text transcription of the call, playable audio of the call, category tag of the call, and time and date of the call,
wherein the display further comprises a button that when pressed initiates output of a visual and audio walk-through of the call so that the audio of the call plays while a status indicator bar travels along visual waveform and a highlighter highlights the corresponding word text of the call.
2. The system of claim 1 , wherein the means for processing structured features and unstructured data features comprises a call analytics processor.
3. The system of claim 2 , wherein the call records analyzer comprises a call waveform processor is configured to isolate and identify audio waveforms for each participant in the call record.
4. The system of claim 2 , wherein the call records analyzer comprises call transcriber that transcribes the audio data into text data.
5. The system of claim 4 , wherein the call analytics processor comprises a model configured to vectorize the text data, analyze the vectorization and tag the data with a call category label.
6. The system of claim 1 , wherein the structured data comprises one or more of word count, number of back and forth exchanges between call record participants, call duration, time of call, and day of call.
7. The system of claim 1 , wherein the unstructured data comprises one or more of keywords, keyword parts, keyword locations relative to other words, and word patterns.
8. A method for automatic real-time call categorization comprising:
receiving a call record comprising a recording of a call at a call categorization system,
determining whether one or both of call transcription and waveform analysis of the call is required,
in response to a determination that call transcription is required: transcribing the call into text data and storing the transcription with the call record, analyzing the text data by vectorization to determine a call category type of the call record, and tagging the call record with the determined call category type,
storing a plurality of analyzed and categorized call records according to tagged call category, and in response to selection of a stored call record, displaying specific visual data for the call, wherein the specific visual data comprises a visual waveform showing a breakdown of call participants, a text transcription of the call, audio of the call and the category of the call, and
initiating a visual and audio walk-through of the call so that the audio plays while a status indicator bar travels along the waveform including the breakdown of the call participants, and a highlighter highlights the corresponding word in the text transcription.
9. The method of claim 8 , wherein in response to a determination that a waveform analysis of the call is required: isolating and identifying waveforms for each participant in the call in the call record and storing the isolated waveforms for each participant in the call record.
10. The method of claim 9 , wherein the call transcription and waveform analysis are performed in parallel.
11. The method of claim 8 , wherein the text data is analyzed and vectorized via a model.
12. The method of claim 11 , wherein the model is a machine learning model that continuously updates as it identifies trends in structured and unstructured data in call records associated with particular call categories.
13. The method of claim 12 , wherein the unstructured data comprises hundreds of features including keywords, keyword parts, keyword locations relative to other words, and word patterns.
14. The method of claim 13 , wherein the structured data comprises word count, number of back and forth exchanges between call participants, call duration, and time of call.
15. The method of claim 12 , wherein the call categories are customizable to the user of the system, and the machine learning model is trained using sample call record structured and unstructured data.