ReDat TopicTracker
Voice and text analysis module for advanced processing of context and content in interactions using AI.
ReDat TopicTracker, which also includes DataExtractor, is a system superstructure of the TopicDetection software module.
The module expands the existing possibilities of defining templates for the analysis of topics and summaries of spoken and non-spoken (textual) interactions using machine learning.
Voice and text analysis module for advanced processing of context and content in interactions using AI.
Key features
Features
- Defining topic templates using a set of calls.
- Creating definitions of new topics is less time-consuming.
- Option to use already established queries from the TopicDetection module.
- Continuous development of topic detection using machine learning.
- For the process of continuously increasing the accuracy of the results, the knowledge of all users can be used.
How TopicTracker works
The basic language model processes records that contain the given topic as well as records that do not contain the topic.
From them, the so-called Tracker is created, which is subsequently used as a third option for defining a topic in the TopicDetection module (in addition to file definition and text query definition).
Advantages of the module
A new method of creating a model for topic detection that does not require knowledge of a scripting language (unlike the TopicDetection module).
All colleagues listening to the calls can thus „help“ the analyst. In their standard work, they can mark calls that contain topics of interest (for example, by inserting a note) as well as records that do not contain the given topic. The more records that are marked and added to the training selections, the higher the resulting accuracy of the evaluation model will be.
Trackers trained in this way are available by default in the agenda of the TopicDetection module.
Data extractor
The TopicTracker extension also includes the DataExtractor functionality, which creates summaries from selected records. These are subsequently available in the record detail next to the Speech-to-Text transcript.
This is an experimental functionality that brings a taste of a new interface that uses AI to process records and, based on prompts (AI queries), extract data from calls (e.g. contract number, subscription point, etc.).
Try our user interface
There is nothing easier, leave us your contact details and we will create access tailored to you, exactly according to your requirements.
We also believe that sometimes it is better to touch the environment properly.
