How to build a Business Partnerships Assistant

Workflow Overview

1. Files Upload (doc-0: Files)

• Purpose:
  Lets you upload and process company documents (like mission statements or past event reports).

• How it works:
  Uploaded files are processed (including OCR for scanned documents) to extract text for analysis.

2. Summarize Company Interests (llm-0: Interests)

• Purpose:
  Reads your uploaded documents and summarizes your nonprofit’s interests and focus areas.

• How it works:
  Uses an AI model to analyze the text from your files and generate a concise summary of your organization’s key interests and goals.

3. City Input (in-0: City Input)

• Purpose:
  Allows you to enter the city where you want to find nonprofit events and speaking opportunities.

• How it works:
  You provide the city name (e.g., "Austin, TX") as input.

4. Craft Search Query (llm-1: Summary)

• Purpose:
  Uses the summary of your interests and the city to craft a web search query for relevant events and speaking opportunities.

• How it works:

  • Combines the company interests (from step 2) and the city (from step 3).

  • Generates a search query designed to find events and speaking opportunities that match your nonprofit’s focus.

  • This node is configured to use a web search tool to help find real events.

5. Recommend Best-Fit Events (llm-2: Recommendations)

• Purpose:
  Analyzes the search results and recommends the best-fit events and speaking opportunities for your nonprofit.

• How it works:

  • Takes the company interests and the list of events found in the previous step.

  • Uses AI to recommend which events and speaking opportunities are the most relevant and actionable for your organization.

6. Output Results (out-0: Output)

• Purpose:
  Shows the final results to the user.

• How it works:
  Displays the recommended events and speaking opportunities generated by the previous step.

7. Write to Notion Page (action-1: Write to Page)

• Purpose:
  Automatically adds the formatted event recommendations to a Notion page for easy sharing and tracking.

• How it works:

  • Takes the outputs from the three LLM nodes (interests, search summary, and recommendations).

  • Writes this combined content to a specified Notion page, so your team can access and track the recommendations.

How the Data Flows

1. You upload documents →
   AI summarizes your interests →
   You enter a city →
   AI crafts a search query →
   AI analyzes search results and recommends events →
   Results are shown to you and also written to Notion for sharing.

Node & Edge Map

Special Features

• Automated Notion Integration:
  Your recommendations are not just shown in the app—they’re also pushed to a Notion page for easy access and collaboration.

• Web Search Integration:
  The workflow uses a web search tool to find real, up-to-date events and opportunities.