---
title: "MCP Tools Reference"
slug: "aia-mcp-tools"
description: "Reference information about available tools for Kentik's AI Advisor MCP Server, useful for custom integrations."
tags: ["Kentik MCP Server", "AI Integration Tools", "Session Management"]
status: "update"
updated: 2026-11-10T03:42:57Z
published: 2026-11-10T03:42:57Z
canonical: "kb.kentik.com/aia-mcp-tools"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://kb.kentik.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Tools Reference

Kentik’s [**AI Advisor MCP Server**](/v1/docs/ai-advisor-mcp-server) exposes three tools for interacting with [**AI Advisor**](/v1/docs/ai-advisor):

- [**ask_question**](/v1/docs/aia-mcp-tools#askquestion)
- [**ask_followup**](/v1/docs/aia-mcp-tools#askfollowup)
- [**get_session_history**](/v1/docs/aia-mcp-tools#getsessionhistory)

> [!NOTE]
> **Note**: This reference and the [**MCP Payload Reference**](/v1/docs/aia-mcp-payload) are specifically for developers building custom AI agents or integrations with the Kentik MCP Server. Standard Model Context Protocol (MCP) clients like Claude Desktop handle these tools automatically in the background.

## **Rate Limits**

Because the MCP server uses the [**AI Advisor REST API**](/v1/docs/ai-advisor-api), standard [**rate limits**](/v1/docs/ai-advisor-api#rate-limiting) apply. The MCP server automatically handles periodic polling to respect these limits.

> [!WARNING]
> **IMPORTANT:** Use the `ask_followup` tool instead of creating new sessions via `ask_question` whenever possible to minimize API calls and maintain context.

## **ask_question**

Submits a new question to AI Advisor and creates a new conversation session.

**Parameters**:

- `prompt` (string, required): Your question about the network

**Example**:

```json
{
  "prompt": "Show me all devices that are currently down"
}
```

**Returns**:

- AI-generated answer in Markdown format
- Session ID for follow-up questions
- Real-time progress updates during processing

**Typical Processing Time**: 5-30 seconds

## **ask_followup**

Asks a follow-up question in an existing conversation, maintaining context.

**Parameters**:

- `session_id` (string, required): UUID of the existing chat session.

> [!NOTE]
> **Note**: This value is dynamic and must be retrieved from the previous response.
- `prompt` (string, required): Your follow-up question.

**Example**:

```json
{
  "session_id": "019ab80d-f7a1-739b-99e0-803f501fdea9",
  "prompt": "What about their interface utilization?"
}
```

**Returns**:

- AI-generated answer with conversation context
- Updated session information

> [!NOTE]
> **Note**: The `session_id` is provided in the response from `ask_question`. Save it to continue the conversation.

## **get_session_history**

Retrieves the complete conversation history for a session.

**Parameters**:

- `session_id` (string, required): UUID of the chat session

**Example**:

```json
{
  "session_id": "019ab80d-f7a1-739b-99e0-803f501fdea9"
}
```

**Returns**: Formatted conversation history with all questions and answers.