Contents:

• Overview
• AI Features Available
• Access, Consent, and Data Handling
• Feature Specific Information
• AI Assistant for Admins
• AI Assistant for Sales
• BookIt MCP
• Additional Resources

 

Overview

LeanData embeds AI capabilities across its revenue orchestration platform to help administrators understand routing logic, automate decisions inside FlowBuilder, and surface operational insights through conversational assistants. This statement describes what AI features exist, which models and providers power them, examples of processed data, and what controls are available to administrators.

AI Features Available

LeanData offers the following AI-powered capabilities as of the date of this document:

Feature	What it does
AI Graph Summary	Generates a natural language explanation of your FlowBuilder routing graph, breaking down business logic, technical details, and optimization recommendations.
AI Graph Comparison	Shows a structured summary of changes between the current routing graph and a previously deployed version.
AI Inference Node	A FlowBuilder action node that sends a custom admin-configured prompt to an LLM (large language model) and saves the structured response to a LeanData variable for use in downstream routing logic.
AI Assistant for Admins	A conversational AI embedded in the LeanData App that helps administrators investigate audit logs, understand routing behavior, understand their graphs, and make updates to their graphs.
AI Assistant for Sales	A conversational AI built for sales reps, SDRs (sales development reps), and BDRs (business development reps) who use LeanData Journeys. Helps users research accounts, analyze buying group journeys, and identify next steps for active opportunities
BookIt MCP	An MCP (Model Context Protocol) server that makes BookIt scheduling and operational data available in MCP-compatible AI clients such as Claude and ChatGPT. Admins and reps can query meetings, check availability, retrieve booking links, and take scheduling actions from their own AI client.

 

Access, Consent, and Data Handling

Supported AI Providers

LeanData uses third-party AI providers for inference, including Google Gemini, OpenAI, and Anthropic.

Data Handling

LeanData processes data needed to deliver each AI feature (for example, routing graphs, audit logs, CRM record fields, and your AI query content).

The following table outlines the data that may be sent to various providers by feature:

Feature	Example data sent for inference
AI Graph Summary & Comparison	Routing graph structure (nodes/edges/logic) needed to generate a summary or diff.
AI Inference Node	Your prompt template and any dynamic fields you choose to include in the model response that is returned to the workflow.
AI Assistant for Admins	Context needed to answer your request (within your permission scope), which may include audit logs, graphs and other LeanData/CRM data.
AI Assistant for Sales	Context needed to answer your question (within your permission scope), which may include account and opportunity-related data.
BookIt MCP	Scheduling data needed to respond to a tool call (within the connection's Admin or User scope), such as meeting logs, user details, and availability. This data is returned to the AI client you connect, and your client sends it to the model you choose.

Access and Controls

• Opt-in by feature: AI features are off by default. To use a feature, you must opt in to that feature in LeanData admin settings.
• Disable anytime: You can turn a feature off at any time.
• LLM keys (BYOK): LeanData supports two key models for accessing LLMs:
  • LeanData-managed key: For select features, LeanData provides a default API key.
  • Bring Your Own Key (BYOK): Your organization supplies its own API key for the LLM provider. LeanData currently stores one key per provider per org. The same key is used across all AI features that use that provider.
  • Note: BYOK keys are configured in the LeanData App under Integrations. Customers may select different providers for different features. Keys are securely stored and cannot be retrieved from the UI once applied.

 

Feature Specific Information

Key Management

Feature	LD Managed Key	BYOK
AI Graph Summary & AI Graph Comparison	Yes (default)	Optional
AI Inference Node	No	Required
AI Assistant for Admins	Required	No
AI Assistant for Sales	Required	No
BookIt MCP	No	No

 

AI Graph Summary & AI Graph Comparison

Both features are powered by Google Gemini. LeanData provides a managed Gemini API key by default. Customers may optionally supply their own Gemini key (BYOK). LeanData selects the default model; BYOK customers may request a different supported model.

Data classification

• Example Usage Data: number of summaries generated, number of comparisons run, response latency
• Example Customer Data: routing record, routing graph content, generated summary text, comparison results
• Example Customer Data shared per-instance for debugging: record information explicitly shared by the customer for issue investigation. Retained only as long as needed to resolve the issue.
• Example Derived Anonymous Data: anonymized routing patterns and benchmarks produced from Customer Data after feature opt-in

Supported Models

Provider	Model
Google Gemini	Gemini 3.7 Flash

---

AI Inference Node

The AI Inference Node operates on a BYOK basis only. Administrators select a specific model per node from the models available through the authorized provider's API.

Data classification

• Example Usage Data: number of node executions, response latency, error rates
• Example Customer Data: admin-configured prompt template, dynamic record field values sent to the LLM, the LLM response saved to the LeanData variable
• Example Derived Anonymous Data: prompt patterns and response classifications produced from Customer Data after opt-in

Supported Models

Provider	Model
Google Gemini	Gemini 2.5 Flash
Google Gemini	Gemini 2.5 Flash-Lite
Google Gemini	Gemini 2.5 Pro
Google Gemini	Gemini 3 Flash Preview
Google Gemini	Gemini 3 Pro Preview
Google Gemini	Gemini 3.1 Pro Preview
Google Gemini	Gemini 3.1 Pro Preview Custom Tools
Google Gemini	Gemini 3.1 Flash Lite Preview
Google Gemini	Gemini Flash Latest
Google Gemini	Gemini Flash-Lite Latest
Google Gemini	Gemini Pro Latest
OpenAI	gpt-5.4
OpenAI	gpt-5.4-mini
OpenAI	gpt-5.4-nano
OpenAI	gpt-5.2
OpenAI	gpt-5.1
OpenAI	gpt-5
OpenAI	gpt-5-mini
OpenAI	gpt-5-nano
OpenAI	gpt-4o
OpenAI	gpt-4o-mini
OpenAI	o4-mini
OpenAI	o3
OpenAI	o3-mini
OpenAI	o1
Anthropic	Opus 5
Anthropic	Sonnet 5
Anthropic	Fable 5
Anthropic	Opus 4.8
Anthropic	Opus 4.7
Anthropic	Sonnet 4.6
Anthropic	Opus 4.6
Anthropic	Opus 4.5
Anthropic	Haiku 4.5
Anthropic	Sonnet 4.5
Microsoft Azure	DeepSeek V4
Microsoft Azure	Cohere Command A
Microsoft Azure	SpaceX Grok 4.1
Microsoft Azure	Anthropic Claude Haiku 4.5
Microsoft Azure	OpenAI GPT 5.4

---

AI Assistant for Admins

A conversational AI embedded in the LeanData App that helps administrators investigate audit logs, understand routing behavior, understand their graphs, and make updates to their graphs.

Data classification

• Example Usage Data: frequency chats are started, number of messages per conversation, number of conversations, return-to-previous-conversation events, page URL, average messages per conversation, average conversations, thumbs-up/thumbs-down ratings
• Example Customer Data: individual message content, audit logs used to generate responses, thumbs-down comment text, error report content
• Example Customer Data shared per-instance for debugging: conversation content explicitly shared by the customer for issue investigation. Retained only as long as needed to resolve the issue.
• Example Derived Anonymous Data: anonymized conversation patterns and audit log query patterns produced from Customer Data after opt-in

Supported Models

Provider	Model
Google Gemini	Gemini 3 Flash Preview

---

AI Assistant for Sales

The AI Assistant for Sales is available to users of LeanData Journeys and runs on a LeanData-managed key.

Data classification

• Example Usage Data: number of account research queries, number of journeys analyzed, session duration, response latency, thumbs-up/down ratings
• Example Customer Data: account and buying group records, query content, research outputs
• Example Customer Data shared per-instance for debugging: conversation content explicitly shared by the customer. Retained only as long as needed to resolve the issue.
• Derived Anonymous Data: anonymized account research patterns produced from Customer Data

Provider	Model
Google Gemini	Gemini 3.5 Flash

---

BookIt MCP

BookIt MCP connects MCP-compatible AI clients, such as Claude, ChatGPT, Cursor, or a custom agent, to BookIt. Users authenticate through Salesforce login or a One-Time Code, and the connection's scope (Admin or User) determines which capabilities are available.

BookIt MCP does not use a LeanData-managed key or BYOK. Your organization and the AI client you choose control model selection, prompting, and guardrails. If your client requires an API key for an LLM provider, you configure that key in your client environment. BookIt MCP does not accept or store LLM provider API keys, and token usage and LLM costs are governed by your agreement with your AI provider.

Data classification

• Example Usage Data: tool calls, which are logged server-side for auditability
• Example Customer Data: meeting logs, user details, availability, scheduling links, and pool and meeting type details returned in response to tool calls
• Credentials: data surfaced through MCP does not include raw OAuth tokens or other credentials. Credentials are encrypted at rest using AES-256-GCM.

Supported Models

BookIt MCP works with any MCP-compatible client. The model is determined by the client and provider your organization chooses.

 

Additional Resources

• AI Governance & Trust Overview

For questions or additional assistance please contact LeanData Support.