The flow operates through a series of connected components that handle data retrieval, content generation, and document assembly. The process begins when triggered by a webhook from a CRM stage change or through direct input via the chat interface.

Data collection starts with API request nodes that fetch customer information, deal details, and pricing data from external systems. The system can also incorporate brand guidelines and standard terms through file reading capabilities. For enhanced content relevance, the flow uses vector stores with embeddings to perform retrieval-augmented generation (RAG), pulling relevant product sheets, case studies, and sales collateral based on the specific opportunity.

Content generation relies on structured prompt templates that provide consistent formatting and variable injection points. Language models generate narrative sections while structured output components ensure reliable JSON formatting for elements like line items and contract terms. This structured approach allows downstream components to process and format data consistently.

The assembly phase uses parser components and data operations to shape information into proposal sections. The system can output to various formats including Markdown, text, JSON, or spreadsheets, with options for local storage, cloud storage, or direct integration with document platforms.

Final delivery can route through multiple channels. The system can automatically create and send proposals through document generation and e-signature platforms via API requests, while simultaneously providing previews through chat outputs for immediate review.

The flow can be extended using additional Langflow components for more sophisticated scenarios. Conditional logic can route different proposal templates based on deal size or customer segment. Google Workspace integration enables direct document creation and collaboration. Multiple flows can be chained together where one flow enriches customer data, another assembles the proposal, and a third handles CRM updates and document archiving, all orchestrated through the flow execution API.