Topics are the key to understanding your customer’s satisfaction by automatically labeling and categorizing incoming written feedback.
Clustering and analysis of feedback is fully automated
Easily see and filter for particular Topics
Prioritize Topics based on how often they appear
Track the NPS development per Topic
Direct feedback to the right people in your organization
The Topics Processing Flow
Language detection → The comment’s language is detected for further processing and labeling.
Parser → The comment is checked for grammar and analyzed for meaning. In other words, the parser digests the comment and extracts the message
Stopwords are removed → AI removes unnecessary stopwords (aka filler words) but only to a degree that ensures the meaning of the comment is kept
Stemmer → The stemmer further simplifies the comment by bringing the words to their base form. For example, delivery or delivered becomes deliver
Keyword matching → The keywords from your Topic Dictionary are matched to the keywords in the comment
Labeling → When matched keywords are identified, the comment is labeled with the Topic
Sentiment Analysis → Once the labels are assigned, the AI checks the comment for mood and tonality. In other words, it will detect whether a statement about a Topic is negative, positive or neutral.
With Topics, you can gain more context to issues, understand the relevant and send more targeted Act Workflows.
To learn how to set-up your individualized Topic architecture, see “Configuring your Topics”
During processing, Topics are matched and applied using Keywords.
To learn how to create an effective keyword set, see Keywords.
The sentiments of Topics indicate how your customers feel about each individual Topic within the comment.
The Topics can carry a positive, negative or neutral sentiment.
For more information, see How Does the Sentiment Analysis Work?
Leverage your Topics to identify pitfalls along your customer journey and route feedback to the correct stakeholders.
To learn more, see Using Topics.