Classifier is a part of Contact Center AI, a tool for classifying user dialogues using AI based on its instructions.
Instruction is a text template with requirements for the AI.
For example, the following scenario can be worked out:
- Analyze the brief content of the dialogue and determine whether a potential or existing customer is calling.
- Determine the type of customer appeal: Interest in a new product, Complaint about the quality of services, Appeal in connection with non-working services, Other.
- Determine the degree of the dialogue. Was it a conversation in raised tones, did the customer experience anger or dissatisfaction? Did the customer calm down after talking to the manager?
To implement such a scenario, you need to create three AI instructions. Instructions are located on the Classifier tab.
To add rules by which AI will analyze the dialogue, click Add instruction.
Specify:
- Instruction title — template topic;
- Instruction description — rules by which AI analyzes the dialogue;
- Comment — notes, reminders or instructions;
- Model:
- gpt-3.5-turbo — fast, cheap and accessible AI model. Maximum number of tokens — 4096.
- gpt-4-1106-preview — smarter, contains newer knowledge, costs more. Maximum number of tokens 4096.
- gpt-4o — higher conversation processing speed and lower token consumption reduce the cost of using the model.
- Gpt mode — manual or auto.
- Tokens — selection of the number of tokens for analysis.
- Temperature— from 0 to 1, recommended value - 0.7.
- 0 – responds with the most likely answers and constructions, less creativity.
- 1 – more creativity, unpredictability.
Click Save.
✅ Note: Dialogue decryption reports are located in the Speech Analytics section. |
To manage the instruction, click on the button with three dots in the upper right corner of the instruction. Instructions can be copied, edited, added to favorites, moved to the archive and moved to group instructions. Grouped instructions are located in the left menu.
✅ Note: Group instructions are not available for editing and deleting. |
Also see:
Contact Center AI >> Speech Analytics
Contact Center AI >> Templates
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article