Workplace AI Automation

Guest feedback,
automated

An AI-powered pipeline that replaced manual weekly reporting with structured, actionable guest feedback analysis — built for a real team, used every week.

RoleDesigner, Builder, Trainer
ContextHoley Moley Golf Club (Fun Labs)
ToolsChatGPT, n8n, Relevance AI
StatusOperational
Read Case Study ↓
scroll
01 — Overview

The Problem

Every week, someone had to manually sift through guest feedback data — pulling CSV exports from ReviewTrackers, reading through reviews, and writing up a structured report for the team. It was tedious, inconsistent, and easy to deprioritize when things got busy.

Meanwhile, the feedback itself was valuable. Guests were telling us exactly what was working and what wasn't. We just weren't processing it fast enough to act on it.

⚠ The Gap
Manual reports took 90 min/week and were inconsistent
Someone had to read every review, identify themes, and write a summary by hand. Quality depended on who did it and how much time they had.
✓ The Fix
Automated pipeline processes 3x the reviews in a third of the time
AI-powered automation handles the full cycle — from CSV ingestion to structured report — consistently, every single week.
67%
Reduction in manual reporting time
90→30
Min weekly report time
3x
Review volume (5/week → 15/week)
3.6→4.5
Google rating improvement
02 — What I Built

Two connected systems

Two connected systems that transformed how the team processes guest feedback.

🤖
AI Feedback Summarization Tool
ChatGPT-based tool. Staff paste reviews, get structured summaries — themes, sentiment, callouts — in seconds. No technical knowledge required, just paste and go.
ChatGPT Custom Prompts
⚙️
ReviewTrackers → Report Pipeline
n8n + Relevance AI automation. Takes the weekly CSV export from ReviewTrackers, auto-generates a formatted report with themes, sentiment breakdown, and flagged outliers.
n8n · Relevance AI
Step 1
ReviewTrackers CSV exported
Weekly feedback data is exported from ReviewTrackers as a structured CSV file containing all new guest reviews.
Step 2
n8n processes and routes data
The automation workflow ingests the CSV, cleans the data, and routes it to the AI analysis layer for processing.
Step 3
Relevance AI analyzes feedback, identifies themes, flags outliers
AI models parse each review for sentiment, extract recurring themes, and flag any outlier feedback that needs immediate attention.
Step 4
Structured report generated and delivered
A formatted, consistent weekly report is automatically generated and delivered to the team — ready to act on without any manual work.
03 — Impact

The numbers that matter

67%
Reduction in manual reporting time
90→30
Minutes per weekly report
3x
Review volume processed weekly
3.6→4.5
Google rating improvement
04 — Reflection

Real work, real impact

This wasn't a side project — it was a real operational problem with a real solution that people actually use. Building it required understanding the workflow, designing for non-technical users, and making something reliable enough that the team would trust it.

"The best automation solves a problem people feel every week, not a hypothetical one."
— Andre Espinoza
🎓
Training is design
Training matters as much as building — adoption is a design problem. If the team doesn't understand it, they won't use it, no matter how well it works.
🔧
No-code is real
AI + no-code tools (n8n, Relevance AI) make operational automation accessible without a dev team. You don't need engineers to solve engineering-shaped problems.
💼
Work ≠ portfolio
Doing this at work, not just in a portfolio project, changes how you think about reliability and user trust. Real stakes sharpen every decision.

Want to see what automation can do?

This project turned messy guest feedback into structured weekly insight. Let's talk about what your data could look like.

Get in Touch ← Back to Portfolio