Success Story - en

Clusterization of Contact Center dialogues

Description:  building a system for automatic processing of incoming calls to the contact centre assumes the presence of a specified taxonomy, which will be used to categorize the request and then process it. When working with a large number of contact centres, you need a system for quickly analyzing dialogues and then building a taxonomy of categories for a specific contact centre.

Context:  it is important for CC analysts to quickly understand what topics are included in the dialogue corpus in order to quickly automate their work. Manually dividing such cases into topics is a labour-intensive task.

Decision:  a system of hierarchical clustering of contact centre dialogues is proposed, which allows creating a taxonomy of intents. 

Integration into the BI Department for automation of customer contact centres:
  • Reducing the load on the analyst to allocate automation classes by 80%.

Harmonization of taxonomy:
  • Reducing the time to identify a new category by 60%;
  • Allocation of new intents in the flow of requests with a quality of 70%.

The selection of entities and the analysis of interpretability:
  • Selecting named entities and filling in the client card is 10% more accurate;
  • Increase in the share of interpreted intents relative to the old model by 40%.

To build the model, we used:
  • Dialogues for multiple contact centres;
  • Assessor markup of paraphrases for each dataset.

Simulation result:
  • Model for soft hierarchical clustering of dialogues;
  • Final taxonomy of categories with a description;
  • A server-side custom application with an API interface.

Customer: Contact Center, Telecom

Technology stack: TopicNet, BigARTM, Flask, Python, PyTorch. 
Natural Language Processing Contact Centers Research Division Engineering Division