Contents:

• Overview
• Example Use Cases
• Enabling AI Graph Summary
• Generating an AI Graph Summary
• Interpreting Graph Summaries
• Working with Summaries
• Permissions and Access
• Using Your Own API Key
• FAQs

 

Overview

LeanData's AI Graph Summary feature uses artificial intelligence to generate natural language explanations of your FlowBuilder Routing Graphs. This powerful tool automatically analyzes your routing logic and creates summaries that explain what your graph does, how it works, and where there may be opportunities for optimization.

AI Graph Summary helps you understand complex routing configurations by breaking them down into clear business logic, technical details, and actionable recommendations. Whether you're onboarding new team members, troubleshooting routing issues, or optimizing your processes, AI Graph Summary provides the insights you need in an easy-to-understand format.

Example Use Cases

• New Administrator Onboarding: Quickly understand inherited routing graphs without extensive documentation or tribal knowledge.
• Cross-Team Collaboration: Share clear explanations of routing logic with Sales, Marketing, and leadership teams.
• Troubleshooting: Identify potential issues, gaps, or redundancies in your routing configuration.
• Process Optimization: Discover opportunities to improve routing efficiency based on AI-generated recommendations.
• Documentation: Create and maintain up-to-date documentation of your routing processes.
• Knowledge Transfer: Preserve institutional knowledge when team members transition roles.

 

Enabling AI Graph Summary

LeanData administrators can control access to AI functionality from LeanData Admin Settings.

• To enable, navigate to Admin > Settings > AI Tools in the LeanData app.
• Toggle on the AI Graph Summary and AI Graph Comparison setting.

By default, LeanData uses its own Gemini key as the LLM provider. If you prefer to use your own Gemini key, you can provide it in Integrations. See the LeanData Admin - Using Your Own Gemini API Key for more information.

 

Generating an AI Graph Summary

From the LeanData App navigate to Routing > [Object] > FlowBuilder, then open a routing graph. You can generate summaries for drafts as well as deployed graphs.

From within a FlowBuilder graph, click the AI Graph Summary button in the upper left toolbar to generate a summary of the currently displayed graph.

If generating a summary for the first time, you can designate if you wish to receive an update in the pop-up modal.

Depending on the complexity of your graph, summary generation can take up to several minutes. You may continue to work on your FlowBuilder graph while the summary generates.

Once complete, the summary will be displayed in a panel with multiple sections:

 

Interpreting Graph Summaries

AI-generated summaries follow a consistent, structured format designed to serve both technical and non-technical audiences. Each summary includes the following sections/tabs:

Summary Tab

The Summary tab provides a high-level overview of the graph's purpose, business outcome, and key structure. This section is written for non-technical stakeholders such as Sales, Marketing, and leadership teams who need to understand what the graph accomplishes without diving into technical details.

 

Business Logic Tab

The Business Logic section outlines more detail, breaking down the graph into sections organized by color and node types and plain language description behind the intent of these sections. This tab may include:

• Flow Name & Purpose: A clear description of what the flow is designed to accomplish.
• Graph Groups: Implied or actual graph groups, including their colors and names (if applicable).
• Entry Criteria: What triggers a record to enter the flow.
• Routing Rules: Description of how records are routed by geography, product, named representative, or channel.
• Matching Rules: Objects used for matching and simplified logic (e.g., email match, domain match).
• Conversion & Assignment: When and how leads convert and how ownership is assigned.
• Alerts & Notifications: Any outbound alerts or notifications configured in the flow.

This section also highlights strengths and weaknesses in your business logic, such as robust duplicate detection or missing fallback handling.

Technical Logic Tab

The Technical Logic section provides detailed information for Administrators, Developers, and QA teams. This section includes:

• Entry Node Configuration: Triggering object and fields used to initiate the flow.
• Decision Nodes: Detailed descriptions of all branching logic.
• Match Nodes: Criteria, thresholds, fallback behavior, and tie-breakers.
• Update Nodes: Fields updated, default values, and logic used.
• Convert/Assign Nodes: Logic for record conversion and assignment.
• Alert Nodes: Templates, channels, and conditions for notifications.
• Error Handling & Fail Paths: Behavior when errors occur, including common failure modes.

Throughout this section, the AI embeds contextual and actionable recommendations for improving your technical implementation.

Recommendations Tab

The Recommendations Tab will give a snapshot of the strengths of your graph, highlight gaps and risks, and provide recommendations.

 

Working with Summaries

Editing Summaries

In the Summary tab, click the pencil icon in the upper right to edit your generated summary. This allows you to add your own details or notes and save it to preserve your edits.

Click the checkmark icon once you are finished with your edits to save them.

Saved summaries are associated with your routing graph and can be accessed anytime from the AI Summary interface.

Regenerating Summaries

When you make changes to your routing graph, you can regenerate the summary to reflect those updates:

To regenerate a summary, click the Regenerate icon in the upper right of your summary panel.

The system will analyze the current state of your graph and generate an updated summary.

If you have a previously saved summary, the AI will attempt to preserve consistent tone and structure while incorporating the new changes.

Exporting Summaries

To export the current summary, click the external link icon to open it in a separate tab.

On the separate tab, you can use your browser's Print to PDF or share capabilities to save or share the summary.

Exported PDFs maintain all sections, tables, and formatting from the original summary, making them ideal for documentation, presentations, or team handoffs.

 

Permissions and Access

Access to AI Graph Summary functionality is controlled by LeanData permission sets:

Permission Set	Access Level
LeanData Dashboard Access	No access
LeanData Read Only with Routing Insights	Read-only access to routing graph summaries
LeanData Custom Objects Full Access	Ability to create and read routing graph summaries
LeanData View	No access

 

Using Your Own API Key

LeanData AI Graph summaries use a standard LeanData Gemini API key to generate summaries. If you would prefer, you can supply your own Gemini API Key. However LeanData will still select the model that processes your graph summary.

To use your own Gemini API Key, please follow the instructions given in the LeanData Admin - Using Your Own Gemini API Key guide.

Once authorized, LeanData will use your key when prompting Gemini within LeanData AI tools.

 

FAQs

Can I generate summaries for invalid or incomplete graphs?

• Yes. AI Graph Summary can generate summaries for incomplete or invalid graphs. The summary will include validity warnings and still provide analysis of the configured logic.

How long does it take to generate a summary?

• Summary generation time will vary depending on the complexity and size of your routing graph. It can be as quick as several seconds, but may take up to several minutes.

Will my edited summaries be overwritten when I regenerate?

• LeanData AI Summary will attempt to preserve consistent tone and structure when regenerating summaries based on previous versions. However, significant graph changes may result in substantial rewrites. It's recommended to save important custom documentation separately if needed.

Is my routing data sent outside of LeanData for AI processing?

• Data is only analyzed during processing, and no personally identifiable information (PII) is included in the data sent for AI analysis.

What AI technology powers this feature?

• AI Graph Summary uses Google's Gemini Flash model to analyze routing graphs and generate natural language summaries. LeanData chooses the model behind AI Graph Summary and may change it as necessary, for example to improve performance or move to a newer model.

What if my organization has concerns about the use of AI?

• AI Graph Summary may be disabled for organizations with AI sensitivity concerns. To disable, from the LeanData app, navigate to Admin > Settings > AI Tools.

 

For questions or additional assistance please contact LeanData Support.