Google Drive Integration

Connecting Google Drive lets ScaleFlow access only the files you select through Google Picker. You can:

Sync documents into Knowledge so AI can read content (FAQs, catalogs, policies).
Attach Drive to an AI Agent so the agent can list files and view metadata through MCP tools.

Use this integration when you want to:

Let your team store documents on Drive and have Smart Assistant answer based on those files.
Let agents find the right file or folder on Drive in operational workflows.
Avoid granting access to all of Drive — only explicitly selected files are opened.

Google Drive vs Google Sheets

	Google Drive	Google Sheets
Primary purpose	Documents (PDF, Word, and more)	Spreadsheets, row/column data
Read content for AI	Through Knowledge (file sync)	Through MCP `read_rows` or Knowledge
MCP tools	`list_files`, `get_file` (metadata)	`read_rows`, `append_row`, and more
Dedicated doc	This guide	Google Sheets Integration

Before You Start

Prepare the following:

A Google or Google Workspace account with the files you need.
Permission to manage integrations in ScaleFlow.
Files already uploaded to Google Drive (PDF, DOCX, TXT, and more).
About 5-10 minutes for the first connection.

Access scope (drive.file): ScaleFlow does not see your entire Drive. It can only access files you select through Google Picker when creating Knowledge or configuring an agent. Files that were not selected cannot be read.

Step Overview

Step	Where	What to do
1	ScaleFlow	Integrations -> Google Drive -> Connect
2	Google	Sign in and Allow access
3	ScaleFlow	Run Test connection and confirm Connected status
4	ScaleFlow	Create a Knowledge source from Google Drive or attach Drive to an AI Agent
5	ScaleFlow	Select files through Google Picker -> sync / test

Step 1 - Connect Google Drive in ScaleFlow

Open the left menu -> Integrations.
Find the Google Drive card -> click Connect.
On the setup screen, review Permissions required:
- Access only the files and folders you choose in Google Picker — only the files you select for automation and knowledge sync.
Click Continue with Google.
Select your Google account and click Allow.
After authentication, the connection appears with Connected status.

Step 2 - Manage the Connection

Open Integrations -> Google Drive and select a connection.

Display tab

Connection name — rename the connection (for example: Google Drive — Marketing, Google Drive — Support).
Google account metadata (read-only): Account name, Google email, Connected on.

Google Drive Display tab

AI and Automation tab

An overview of Drive workflows with AI.
Google Drive actions — MCP tools.
Recommended setup flows: Document Discovery, Metadata Sync, Knowledge Source Sync.

Google Drive AI and Automation tab

Quick actions

Action	When to use it
Test connection	After connecting, or when you suspect the token expired
Reconnect	Google revoked access
Disconnect / Delete	Stop or remove the connection

Step 3 - Use Google Drive as Knowledge (Read Document Content)

This is the most common approach: AI reads file content (FAQ, catalog) through Knowledge, not through Drive MCP.

Open AI -> Knowledge -> Add knowledge.
Select Source type: Google Drive.
Choose the Google Drive connection (with Connected status).
Click Select files and choose one or more files through Google Picker.
Click Add knowledge.
On the Knowledge detail page, click Sync knowledge / Synchronize so ScaleFlow ingests the content.

Knowledge Google Drive source

Attach the Knowledge source to an AI Agent (Capabilities tab -> Knowledge).
The agent uses the queryKnowledge tool to answer based on synced documents.

Step 4 - Attach Google Drive to an AI Agent (MCP)

Use this when the agent needs to find files or view metadata (name, type, last modified date, link) — not read PDF/DOC content directly through MCP.

Open AI -> Agents and select an agent.
Open the Capabilities tab -> add the Google Drive integration.
Click Select files and choose files or folders through Google Picker.
Click Save Draft -> Test Version.

Drive MCP does not replace Knowledge: if you want the agent to answer from document content, use Knowledge (Step 3). MCP only supports discovery and metadata.

Test Google Drive files

Step 5 - Sample Runbook to Test the Agent

Paste the following into the Runbook field (after selecting files in Capabilities):

#GoogleDrive: You can look up Google Drive through google_drive_list_files and google_drive_get_file.

## Rules
- When the user asks "what files are there" or "find a document" -> use google_drive_list_files (pageSize: 10).
- When the user asks for details about a file -> use google_drive_get_file with fileId.
- When the user asks for DOCUMENT CONTENT (FAQ, policy) -> use queryKnowledge from attached Knowledge, do NOT guess from Drive metadata.
- Do not invent file names or links.

## Test prompts
- "List the Drive files I connected"
- "When was [file name] last modified?"

Suggested test prompts:

Goal	Question
List files	"Show me the list of Google Drive files I connected."
Metadata	"What type is [file name] and what is the view link?"
Document content	"According to the FAQ document, what is the return policy?" -> requires synced Knowledge

MCP Tools - What Can an Agent Do?

Google Drive MCP actions

List Files (`google_drive_list_files`)

Lists files and folders the connection can access.

Parameter	Description
`query`	Search query (for example: `name contains "FAQ"`). Default: `trashed = false`
`pageSize`	Number of results (default: 10)
`pageToken`	Pagination token for the next page

Returns: files (id, name, mimeType, size, modifiedTime), nextPageToken.

Get File Metadata (`google_drive_get_file`)

Returns metadata for one file or folder.

Parameter	Required	Description
`fileId`	Yes	File ID on Drive

Returns: id, name, mimeType, size, modifiedTime, webViewLink, iconLink.

There is no MCP tool to read file content. To let AI read PDF/Word files, use Knowledge sync.

Get the File ID from Google Drive

Open the file in your browser. The URL usually looks like:

https://drive.google.com/file/d/1ABC...xyz/view

The part between /d/ and /view is the fileId.

In most cases, the agent gets fileId from List Files results — you do not need to copy it manually.

Recommended Workflows

1. Document Discovery Assistant

The agent calls List Files to find files by name or conversation context -> returns the webViewLink to the user.

2. Document Metadata Sync

The agent reads Get File Metadata (last modified date, file type) -> records it in a ticket or another sheet.

3. Knowledge Source Sync

An admin selects FAQ files or folders on Drive -> creates Knowledge -> syncs periodically -> Smart Assistant answers through queryKnowledge.

Real-World Workflow

An admin connects Google Drive (Marketing account).
The admin creates Knowledge Gym FAQ — selects 5 PDF files from Drive -> Sync knowledge.
The admin creates a Gym Sales agent, attaches Knowledge and (optionally) the Drive integration.
The agent answers product questions from Knowledge; when a source file link is needed, it uses MCP metadata.
When a new FAQ file is added -> add it to Knowledge and sync again.

Quick Troubleshooting

Google Picker does not open

Confirm the Drive connection is Connected.
Try Reconnect and accept permissions again.
Check that the browser is not blocking popups.

Agent/Knowledge cannot see a file

The file must be selected through Picker — not every file on Drive is visible automatically.
Select the file again in Knowledge or agent Capabilities.

Knowledge sync fails

The file may be too large or in an unsupported format — try PDF/DOCX/TXT.
Access to the file was lost -> Reconnect and select the file again.

MCP list files returns empty

No files were selected in agent Capabilities, or no files have been opened through Picker yet.
Try a specific query, such as name contains "FAQ".

Refresh token error

Go to Google Account permissions, remove ScaleFlow access, then Reconnect.

Agent answers document content incorrectly

Drive MCP only provides metadata — you need synced Knowledge and Instructions that use queryKnowledge.

Best Practices

Name connections by team: Google Drive — Support.
Split Knowledge by topic (FAQ, Pricing, Policy) instead of one oversized knowledge source.
Re-sync Knowledge after updating files on Drive.
Use Drive MCP for finding files / metadata; use Knowledge for reading content.
Select only the files you need — this reduces risk and makes access easier to audit.

Suggested Screenshot List

File	What to capture
`google-drive-integration-card.png`	Google Drive card + Connect
`google-drive-setup-permissions.png`	Setup screen + Continue with Google
`google-drive-connections-list.png`	Connected connections list
`google-drive-display-tab.png`	Display tab
`google-drive-ai-automation-tab.png`	AI and Automation tab
`google-drive-agent-picker.png`	Agent Capabilities + Select files
`google-drive-mcp-tools.png`	List Files / Get File Metadata

Reusable images: integration-connection.png, connect-google-drive.png, knowledge-google-drive.png.

Google Drive vs Google Sheets​

Before You Start​

Step Overview​

Step 1 - Connect Google Drive in ScaleFlow​

Step 2 - Manage the Connection​

Display tab​

AI and Automation tab​

Quick actions​

Step 3 - Use Google Drive as Knowledge (Read Document Content)​

Step 4 - Attach Google Drive to an AI Agent (MCP)​

Step 5 - Sample Runbook to Test the Agent​

MCP Tools - What Can an Agent Do?​

List Files (google_drive_list_files)​

Get File Metadata (google_drive_get_file)​

Get the File ID from Google Drive​

Recommended Workflows​

1. Document Discovery Assistant​

2. Document Metadata Sync​

3. Knowledge Source Sync​

Real-World Workflow​

Quick Troubleshooting​

Google Picker does not open​

Agent/Knowledge cannot see a file​

Knowledge sync fails​

MCP list files returns empty​

Refresh token error​

Agent answers document content incorrectly​

Best Practices​

Suggested Screenshot List​

Read Next​