AI Workflow Automation

Problem

As a PM, I was spending too much time manually tracking and interpreting user feedback from platforms like Discord and Slack. This process was not only inefficient and error-prone but also distracted me from focusing on higher-priority, high-impact tasks that required my attention. Valuable insights were often buried in scattered messages, making it difficult to identify trends, prioritize effectively, and share actionable feedback with the team in a timely way

Objective

Design and implement a workflow automation that collects user feedback from multiple channels (e.g., Discord, Slack), interprets it using AI (LLMs), and generates structured, actionable insights on a regular cadence. The goal is to reduce manual triage time, increase visibility into user pain points and requests, and support data-informed product decision-making.

Solution

1. Feedback Collection and Classification Workflow

• Goal: Automate the intake, interpretation, and logging of real-time user feedback

• Key Functions:

  • Collects messages from platforms like Discord and Slack

  • Uses an AI agent node to:

    • Categorize each message into: Bug Report, Feature Request, General Feedback, or Ignore

    • Assign a priority level (High, Medium, Low) based on urgency and product impact

  • Logs structured data into Google Sheets for centralized access

2. Weekly Feedback Digest Report

• Goal: Generate and distribute a weekly summary of key user feedback insights.

• Key Functions:

  • Runs via a cron trigger every Monday

  • Filters and groups feedback based on category and priority

  • Formats summaries using an AI agent node

  • Sends categorized highlights (e.g., top feature requests, critical bugs) to a dedicated Slack channel

3. Interactive Feedback Query Bot (ProductBot)

• Goal: Enable stakeholders to access on-demand feedback insights via natural language.

• Key Functions:

  • Listens for questions in Slack (e.g., “What’s the most requested feature this week?”)

  • Uses an AI agent to interpret queries

  • Searches feedback logs in Google Sheets

  • Generates a tailored summary and replies to the user directly in Slack

Watch the video on how I did this here!