MOM Generator

The AI-powered Meeting Minutes (MOM) generator transforms raw meeting notes into structured, professional documentation.

Overview

ConsultFlow uses OpenAI's GPT models to:

• Extract key information from unstructured notes
• Identify action items with assignees and due dates
• Summarize discussions by speaker
• Format output in professional Markdown

How It Works

Raw Meeting Notes → OpenAI GPT → Structured MOM → Markdown Output

1. User pastes meeting transcript or notes
2. AI processes with specialized prompt
3. Structured MOM generated in Markdown
4. Preview displayed with formatting

Output Format

The generated MOM follows this structure:

# Meeting Minutes

## Overview
[80-word abstract summary of the meeting]

## Discussion Items
[Speaker Name]:
- Point 1 discussed
- Point 2 discussed

[Another Speaker]:
- Their discussion points

## Action Items
| Action Item | Assignee Name | Due Date |
|-------------|---------------|----------|
| Task description | Person Name | 12/25 & Wed |

## Date and Time
[Extracted or TBD]

## Location
[Extracted, Virtual, or TBD]

## Meeting Title
[Extracted or generated]

## Attendees
- Name 1
- Name 2
- Name 3

## Materials Used
[Documents, presentations mentioned]

## Agenda
• Item 1
• Item 2

## Decisions
[Key decisions made]

## Detailed Minutes
### Project Progress Report
[Status updates]

### Confirmation of Issues/To Do Status
[Issue updates]

### Other Contact Items
[Additional discussion]

## Next Meeting
- **Date and Time:** [Extracted or TBD]
- **Location:** [Extracted or TBD]
- **Attendees:** [Expected attendees]

Using the MOM Generator

Standalone Mode

1. Go to MOM Generator from sidebar
2. Paste meeting notes in the text area
3. Click "Generate MOM"
4. Review the output
5. Copy or use in workflow

In Workflow Execution

1. Execute a workflow with "Generate MOM" action
2. Paste notes in the transcript field
3. Click "Generate MOM"
4. MOM auto-fills email body
5. Continue with other actions

Input Tips

Good Input Example

Project Status Meeting - December 26, 2024
Attendees: John Smith (PM), Sarah Johnson (Dev), Mike Brown (QA)

Discussion:
John: Q4 results show 15% growth over Q3. We're on track for annual targets.
Sarah: The new feature deployment is complete. We had some minor bugs but they're fixed.
Mike: Testing is 90% complete. Found 3 critical bugs, all assigned to Sarah.

Action Items:
- John to prepare Q1 forecast by January 5
- Sarah to fix critical bugs by December 28
- Mike to complete final testing by December 30

Next meeting: January 3, 2025 at 2:00 PM
Location: Conference Room A

Tips for Better Results

1. Include attendee names - Helps identify speakers
2. Add dates - For action item due dates
3. Use clear structure - Separate topics clearly
4. Quote speakers - "John said..." or "John: ..."
5. List action items - Even informally mentioned ones

AI Configuration

Environment Variables

# Standard OpenAI
OPENAI_API_KEY=sk-your-api-key
OPENAI_MODEL=gpt-3.5-turbo  # or gpt-4

# Azure OpenAI (alternative)
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com
AZURE_OPENAI_API_KEY=your-api-key
AZURE_OPENAI_DEPLOYMENT_NAME=gpt-35-turbo

Model Comparison

Model	Speed	Quality	Cost
gpt-3.5-turbo	Fast	Good	Low
gpt-4	Slower	Excellent	Higher
gpt-4-turbo	Medium	Excellent	Medium

Prompt Engineering

The system prompt instructs the AI to:

1. Extract meeting metadata
2. Summarize in 80 words max
3. Group discussion by speaker
4. Create action item table
5. Use professional language
6. Format as Markdown

Pre-filling Workflow Data

When MOM is generated, ConsultFlow extracts:

Email Pre-fill

• To: Extracted attendee emails
• Subject: "Meeting Minutes - [Meeting Title]"
• Body: Full MOM content

Calendar Pre-fill

• Title: Next meeting title
• Date/Time: Extracted from "Next Meeting"
• Attendees: Meeting participants

API Usage

Generate MOM Endpoint

POST /mom/generate
Content-Type: application/json
Cookie: session_cookie

{
  "meetingNotes": "Your meeting transcript here..."
}

Response

{
  "mom": "# Meeting Minutes\n\n## Overview...",
  "structured": {
    "meetingTitle": "Project Status Meeting",
    "dateTime": "December 26, 2024",
    "attendees": ["John Smith", "Sarah Johnson"],
    "actionItems": [
      {
        "task": "Prepare Q1 forecast",
        "assignee": "John",
        "dueDate": "January 5"
      }
    ],
    "nextMeeting": {
      "dateTime": "January 3, 2025 at 2:00 PM",
      "location": "Conference Room A"
    }
  }
}

Troubleshooting

"Failed to generate MOM"

1. Check OpenAI API key is valid
2. Verify API key has sufficient credits
3. Check backend logs for specific error

Poor Quality Output

1. Provide more structured input
2. Include clear speaker attributions
3. Try a better model (gpt-4)
4. Add more context to notes

Missing Action Items

1. Explicitly list action items in input
2. Use phrases like "Action: John to..."
3. Include due dates where possible

---

Pro Tip

For best results, include speaker names with colons: "John: I think we should..." This helps the AI attribute discussion points correctly.