As businesses and educational programs scale, evaluating user progress manually becomes a massive bottleneck. Recently, I built an automated workflow to solve this exact problem: a system that instantly analyzes student progress reports using AI and logs customized feedback back into a database—with zero human intervention.
In this post, I will walk you through how I integrated Google Forms, Make.com, Google Gemini AI, and Google Sheets to create an automated AI Career Coach. Whether you want to build this yourself or are looking for an automation expert to streamline your business, this guide will show you what’s possible.
The Tech Stack
Data Entry: Google Forms
Database: Google Sheets
Automation Engine: Make.com
AI Engine: Google Gemini (Gemini 3.8 Flash)
The Use Case
Students in a training program fill out a weekly progress form containing their Name, Career Goal, and Progress So Far.
Instead of an instructor reading each submission, the data is automatically sent to Google Gemini. Gemini acts as an expert career coach, evaluates the student's progress, assigns a status (On Track / Needs Review / Action Required), and writes a summary. Finally, Make.com updates the exact row in Google Sheets with the AI's assessment.
Step-by-Step Architecture
Step 1: The Trigger (Catching the Data)
The first step in Make.com is setting up the trigger.
Pro-Tip: Initially, I used the standard Google Forms "Watch Responses" trigger. However, this trigger only grabs the form answers and doesn't know which row the data lands on in Google Sheets.
To fix this, I used the legacy "Watch Responses in Google Sheets" module. This allows Make to monitor the linked spreadsheet directly, catching the form answers and generating a crucial Row number data variable needed for the final step.
Step 2: Integrating Google Gemini AI
Next, I connected the Google Gemini AI module (Generate a response). I mapped the dynamic data from the form into a custom prompt.
Here is the exact prompt structure I used:
You are an expert student career coach and evaluator.
Evaluate this student's progress update:
- Student Name: {{Full Name}}
- Career Goal: {{Ultimate Career Goal}}
- Progress So Far: {{Progress Update}}
Provide a concise assessment in exactly this format:
Status: [Choose one: On Track / Needs Review / Action Required]
Summary: [1-2 sentences evaluating their progress towards their goal and any suggested next steps.]
The Result
Once the automation schedule is turned on, the magic happens seamlessly. A student submits a form, and within seconds, their specific row in Google Sheets is updated with a highly tailored, actionable AI evaluation.
This workflow eliminates hours of manual review and provides immediate value to the end user.
Let's Work Together!
Building this pipeline was incredibly rewarding. Automating data entry, integrating AI into daily workflows, and connecting different apps via APIs is where I thrive.
If your business is stuck doing repetitive manual tasks, wasting hours on data entry, or you want to integrate powerful AI models like Gemini or OpenAI into your systems—I can help.
Let's automate your business.
Drop a comment below, or reach out to me directly to discuss how we can build custom workflows that save you time and scale your operations.
Top comments (2)
I chose Make.com to orchestrate this instead of writing a custom Python script so I could deploy it in hours instead of days. For the developers reading this: when building AI pipelines, do you prefer visual tools like Make/Zapier, or do you strictly write custom code? Curious to hear your trade-offs.
The trickiest part of this build was engineering the Gemini prompt so it output the exact format my database needed, without adding extra conversational fluff. Has anyone else here struggled with getting LLMs to return perfectly structured data? What are your favorite prompting tricks?