Assistants in SpeechSense

An assistant is an AI assistant created in SpeechSense. It can automatically analyze dialogs and complete various tasks based on the provided instructions. Assistants are part of the SpeechSense resource hierarchy and are created at the project level.

Basic assistant settingsBasic assistant settings

When creating an assistant, you need to provide the following settings:

• Instructions: Plain text description of your assistant’s behavior and tasks.

  Instructions must include the request context and the task for the model. The more specific your instructions, the more accurate the model's output.

• Model: LLM to use for the assistant.

• Output: Output fields to write the extracted information to.

You can also set dialog filtering conditions and test the assistant before you start using it.

Assistants have restrictions on the number of characters. Learn more in Quotas and limits.

Knowledge baseKnowledge base

You can add a knowledge base as an additional tool to enhance your assistant’s capabilities. The assistant will use the knowledge base information to complete the instructions.

SpeechSense supports uploading PDF files for the knowledge base. The following restrictions apply:

• Maximum file size: 30 MB.
• Maximum number of files in one knowledge base per assistant: 10.
• You cannot edit a knowledge base.

You can use the same knowledge base with multiple assistants within a single space.

Dialog analysisDialog analysis

Assistants are designed to analyze dialogs in a project based on the instructions you provide. An assistant can help with the following tasks:

• Extracting key information from dialogs.
• Classifying dialogs by different criteria.
• Generation or calculation of metrics.
• Detection of behavior patterns in customers and agents.

Using an assistantUsing an assistant

To use an assistant in your project, you need to activate it manually. You start paying for the assistant as soon as you activate it.

AdvantagesAdvantages

1. Automated analysis.

   Reduced manual effort when analyzing large amounts of dialogs.

2. Flexible configuration.

   Performance tailored to various business tasks using instructions.

3. AI integration.

   Use of state-of-the-art models for deep insights into the dialog context.

To create, update, or delete an assistant, follow this guide.