# Chat Assistant

A Chat Assistant (also known as a virtual assistant or conversational agent) is an AI-driven system designed to engage in conversations with users, providing assistance, answering questions, and completing tasks.


## Guide to Chat Assistant

1. Navigate to the TIR dashboard.
2. From the left-side navigation panel:
   - Locate the **RAG** section
   - Click on **Chat Assistant**
3. On the Chat Assistant listing page:
   - You will see a table listing existing assistants (if any)
   - Locate the **CREATE ASSISTANT** button at the top-right
4. Click **CREATE ASSISTANT** to begin creating a new assistant.


## Assistant Settings

After clicking **CREATE ASSISTANT**, you will be redirected to the Assistant Settings page.

On this page:

1. You will see a multi-step configuration layout (typically shown in a left-side step panel such as:
   - Assistant Settings
   - Prompt Engine
   - Model Settings
   - Summary Preview)

2. In the **Assistant Settings** section:
   - Locate the Knowledge Base selection panel
   - A list of available Knowledge Bases will be displayed

3. Select one or more Knowledge Bases:
   - Use checkboxes or selection controls provided in the list
   - Ensure selected Knowledge Bases share the same embedding model

> **Note:** Knowledge Bases with different embedding models cannot be used together in an assistant. However, you can select multiple Knowledge Bases that share the same embedding model.

4. After selecting the Knowledge Base:
   - Locate the **Next** button (typically at the bottom-right of the page)
   - Click **Next** to proceed to the Prompt Engine


## Prompt Engine

The Prompt Engine in a chat assistant is a crucial component that helps generate and manage the prompts given to language models, guiding them to produce relevant and contextually appropriate responses. It plays a key role in enhancing the quality of interactions between users and the assistant.


### Prompt Engine Configuration Steps

1. After clicking **Next**, you will land on the **Prompt Engine** section
2. The interface contains:
   - A configuration panel with input fields
   - Sliders or numeric inputs for tuning parameters


### Prompt Engine Parameters

#### System

- Locate the **System input field**
- Enter the base instructions that define:
  - Assistant behavior
  - Tone
  - Objectives

The System refers to the foundational setup or instructions that define the assistant's overall behavior and objectives.


#### Similarity Threshold

- Locate the **Similarity Threshold slider or input**
- Adjust the value to define the cutoff for relevance matching

The similarity threshold acts as a cutoff point that dictates whether the input is considered similar enough to the training data or knowledge base to generate a relevant response.


#### Keywords Similarity Weight

- Locate the **Keywords Similarity Weight control**
- Adjust the weight to influence how keyword matches impact retrieval ranking

This weight quantifies how closely related certain keywords in a user query are to keywords associated with potential responses or entries in a knowledge base.


#### Top N

- Locate the **Top N field**
- Enter the number of top results to retrieve (e.g., 1, 3, 5)

"Top N" signifies the selection of the top N responses that have the highest relevance or similarity to a user’s query.


## Model Settings

Model settings in chat assistants refer to the configuration parameters and options that dictate how the underlying language model operates, interacts with users, and processes input.


### Model Configuration Steps

1. Click **Next** to navigate to the **Model Settings** section
2. In this section, you will see:
   - Model selection dropdowns
   - Parameter tuning controls (sliders and inputs)


### Model Selection and Validation

1. Locate the **Model Source selector**
2. Choose either:
   - **GenAI**
   - **Endpoint**

3. If **Endpoint** is selected:
   - Select the **LLM Model** from the dropdown
   - Select the **Model Name**
   - Click the **Validate** button (if available)
   - Ensure validation is successful before proceeding

Users must select both the LLM Model from the Model Endpoint and the Model Name, then validate its availability.


### Key Aspects of Model Settings

#### Model Source

Users can choose the model source as either **GenAI** or **Endpoint**.

- If **GenAI is selected**:
  - Select a predefined LLM model

- If **Endpoint is selected**:
  - Select the LLM Model
  - Select the Model Name
  - Validate availability before proceeding


#### LLM Model

Specifies which language model is being used. Different models may excel in various tasks or domains and differ in their ability to understand context and generate responses.


#### Temperature

- Locate the **Temperature slider**
- Adjust the value:
  - Higher values → more creative output
  - Lower values → more deterministic output

Controls the randomness of the model's output.


#### Top P

- Locate the **Top P control**
- Adjust probability threshold for token selection

A sampling technique that selects from the top P percentage of probable next tokens.


#### Presence Penalty

- Locate the **Presence Penalty control**
- Adjust to reduce introduction of new topics

Discourages the model from introducing new topics not mentioned in the conversation.


#### Frequency Penalty

- Locate the **Frequency Penalty control**
- Adjust to reduce repetition

Adjusts the likelihood of repeated tokens in output.


#### Max Tokens

- Locate the **Max Tokens input field**
- Define the maximum output length

Defines the maximum number of tokens the model can generate in a response.


## Summary Preview

In Summary Preview, you can edit Assistant Settings, the Prompt Engine, and Model Settings.

### Steps:

1. Click **Next** to navigate to **Summary Preview**
2. Review all configured sections:
   - Assistant Settings
   - Prompt Engine
   - Model Settings
3. If changes are required:
   - Navigate back using step navigation
4. Once verified:
   - Proceed to finalize assistant creation


## Manage Assistant

### Overview

After successfully creating the assistant:

1. You will be redirected to the **Manage Assistant** page
2. The page contains:
   - A table listing assistants
   - Columns such as name, configuration, and actions

3. Click on an assistant entry to view details in the **Overview section**


## Chat

Chat functionality allows users to interact with the assistant through natural language.


### Access Chat

1. Navigate to the **Manage Assistant** page
2. Locate the assistant row
3. Click on the **Chat** option associated with the assistant
4. The chat interface will open


### Create Conversations

1. Inside the chat interface:
   - Locate the **Create Conversation** option (usually in the left panel)
2. Click **Create Conversation**
3. A new chat session will be initialized
4. Enter your query in the input field and begin interaction


## API Request

API requests enable interaction between the assistant and external systems.


### Steps to Use API Requests

1. Navigate to the assistant interface
2. Locate the **API Request section**
3. You will see:
   - API request examples
   - cURL commands or structured request formats
4. Copy the required request format
5. Use it in your application or API client to interact with the assistant programmatically


## Actions

### Edit Assistant

1. Navigate to the **Manage Assistant** page
2. Locate the assistant in the table
3. Click the **Edit icon** in the Actions column
4. The configuration page will open
5. Modify:
   - Assistant Settings
   - Prompt Engine
   - Model Settings
6. Save changes to update the assistant


### Delete Assistant

1. Navigate to the **Manage Assistant** page
2. Locate the assistant in the table
3. Click the **Delete icon** in the Actions column
4. A confirmation dialog will appear
5. Click the **Delete button** to confirm

The assistant will be permanently removed.


---