Customer Complaint Analyzer

This is a submission for the Runner H "AI Agent Prompting" Challenge

💬 Customer Complaint Analyzer – End-to-End AI Workflow

What I Built

I created an autonomous AI workflow that reads customer complaints from a Google Sheet, analyzes their sentiment and issue type, and simulates full follow-up actions across documentation, email, messaging, and scheduling.

The agent helps streamline customer support tasks, enabling fast triage of issues without any manual filtering or writing code.

Demo

🎥 Watch the demo here:

👉 [https://drive.google.com/file/d/1-RS9cO3BEkJSWFdKoGpC19qAPJRw470h/view?usp=sharing]

Resources

If the video doesn't load, you can also check the attached PDF summary generated by Runner H.
(RunnerH Output):
https://drive.google.com/file/d/1DYAt5pKtZsy12NxBLVk-LvUJrvg0MSIB/view?usp=sharing
Customer_Complaints_Example_file:
https://docs.google.com/spreadsheets/d/1we19oHz7ALnspENcOxFqdDjLaFF3rsmK/edit?usp=sharing&ouid=113936526163083910773&rtpof=true&sd=true

Prompt File:
https://drive.google.com/file/d/1SKp19IwMrUGgePE5snIp2A0G9GxYe9TY/view?usp=sharing

How I Used Runner H

I used Runner H to perform the following steps in a single prompt:

1. Open and read data from a live Google Sheet containing 10 customer complaints.
2. Analyze each complaint’s sentiment (Positive, Neutral, Negative).
3. Classify the complaint into categories (e.g., Delivery, Product Issue, Payment, Website).
4. Update the sheet with new “Sentiment” and “Category” columns and mark rows as “Processed”.
5. Simulate creating a summary report in Google Docs with totals by sentiment and category.
6. Simulate sending a summary email and posting the results in Slack.
7. Simulate scheduling a meeting in Google Calendar if ≥3 negative complaints were found.

🔁 Workflow: Complaint Processing & Reporting

For each row:

1. Detect sentiment

Options: Positive, Neutral, or Negative

1. Classify complaint type

Options: Delivery, Payment, Product Issue, Website, Other

1. Add two new columns

Sentiment
Category

1. Mark row as "Processed"

After processing all rows:

1. Generate a Google Doc report

Include totals by Sentiment and Category
Add a list of Urgent Issues (Negative + Product Issue)

1. If there is at least one negative complaint:

Schedule a 15-minute meeting with the QA/CX team via Google Calendar

1. Send an email

To: [e-mail]
Content: Summary and link to the report

1. Post a Slack message

Include sentiment/category counts
Attach link to the report

All actions were executed or simulated from a single prompt, showing the power of Runner H as a no-code task delegate.

Use Case & Impact

This type of automation is valuable for QA teams, CX leads, and small businesses dealing with user feedback.

Instead of manually checking each support message, this agent can sort, report, and trigger next steps in minutes—improving team response times, customer satisfaction, and operational efficiency.

Even in simulation mode, this workflow proves that everyday business processes can be handled autonomously with AI agents.

---

Social Love

(www.linkedin.com/in/deivismatheusdccs1234)

---

RunnerH made it incredibly easy to automate what would normally take hours of scripting or manual work. This challenge proves how accessible AI agents can be for solving real-world workflows, no coding required.

Thanks again to the Runner H team for organizing this challenge and inspiring creative agent design...