# Build an AI-Powered Customer Feedback Analysis System using n8n, Jotform, Airtable, and OpenAI GPT ### Introduction & What This Automation Does In this project, we will walk through a **Customer Feedback Analysis System** built using **n8n, Jotform, Airtable, and OpenAI**. The goal of this automation is simple but extremely powerful: **turn raw customer feedback into clear, actionable insights automatically**. Many companies collect feedback through forms, but the process of manually reading every message, understanding the sentiment, and deciding what action to take can be slow and inefficient. Important feedback can easily get overlooked, especially when teams receive large volumes of responses. This automation solves that problem by introducing AI into the workflow. Instead of manually reviewing each feedback entry, the system automatically analyzes the message using AI. It evaluates multiple aspects of the feedback, including: * Sentiment (positive or negative) * Category of the feedback * Urgency level * A short summary of the feedback * Priority score * Confidence score * Root cause of the issue * Recommended recovery direction * A recovery message that can be sent to the customer All of these insights are generated automatically using OpenAI and stored in Airtable alongside the original feedback. This allows teams to quickly understand customer sentiment and prioritize the most important issues without manually reading every message. The automation also ensures that the right people are notified immediately. Depending on the sentiment of the feedback, the system automatically sends notifications through **email and Slack**, ensuring that the operational team can respond quickly when needed. In this project, **Airtable is used as the main data storage**. All raw feedback and AI-generated insights are stored there so the team can easily review and analyze the results. [https://airtable.com/app2B5gVgIrG2Sw89/shrKBDfJ5ggoAnp85](https://airtable.com/app2B5gVgIrG2Sw89/shrKBDfJ5ggoAnp85) Inside the Airtable base, two main tables are used. The **Customer Feedback** table stores all raw feedback data submitted by customers. This includes the customer's first name, last name, email address, the service related to the feedback, the feedback message itself, and the rating from one to five. The **AI Analysis** table stores the AI-generated insights. These include sentiment, category, urgency level, summary, priority score, confidence score, root cause, recovery direction, and recovery message. By structuring the data this way, the team can easily review both the original feedback and the AI-generated insights. If you want to build this automation yourself, you can download the complete n8n workflow JSON from the link below and import it directly into your n8n instance. [https://drive.google.com/drive/folders/1cot945mUUJTYL2Kv86RffznmismtcHB4?usp=sharing](https://drive.google.com/drive/folders/1cot945mUUJTYL2Kv86RffznmismtcHB4?usp=sharing) Since the full workflow is provided, you can explore each node, study how the workflow is built, modify the logic, and adapt the system to your own use cases. Before diving into the workflow explanation, let’s first take a look at the quick demo of how the system works. * * * ### Quick Demo You can watch a quick demo of the entire project below to see the automation working from start to finish. %[https://www.youtube.com/watch?v=l8XI_fTD-Ic] * * * ### Technologies Used and Prerequisites Before building this automation, make sure you have access to the following tools and services. **Technologies Used** * n8n (workflow automation platform) * Jotform (customer feedback form) * Airtable (data storage) * OpenAI GPT model (AI feedback analysis) * Slack (team notifications) * Email service (for sending automated responses) **Prerequisites** To follow this project, you should have: * An **n8n instance** running (self-hosted or cloud) * A **Jotform account** with a feedback form created * An **Airtable base** for storing feedback and analysis * An **OpenAI API key** * A **Slack workspace** with a channel for notifications * An **email account configured in n8n** Once these services are connected inside n8n, the workflow can automatically connect all systems together. * * * ### Workflow Explanation ![](https://cdn.hashnode.com/uploads/covers/682c577bb9129c264d975875/e7bf39fa-cb29-457c-a09c-d235adddab80.png align="center") This automation is powered by a single main workflow called **Customer Feedback Analysis Main Workflow**. This workflow connects Jotform, Airtable, OpenAI, Slack, and email into one automated pipeline. Let's walk through the workflow step by step. **Jotform Trigger** The workflow begins with a **Jotform Trigger**. Whenever a customer submits the feedback form, this trigger automatically activates the workflow. Inside the node, you simply connect your Jotform credentials and select the form that will be used. Once configured, every new form submission will automatically start the workflow. **Set Parameters Node** The next step is a **Set Parameters** node. This node defines reusable variables that are used later in the workflow. In this project, it stores the **operational team email address**. By defining this value once, the workflow becomes easier to maintain because it avoids repeating the same value in multiple nodes. **Create Customer Feedback in Airtable** The workflow then stores the raw feedback data in Airtable using the **Create Customer Feedback** node. All form fields from Jotform are mapped directly into Airtable fields. This ensures that every submission is stored consistently inside the **Customer Feedback** table. At this point, the raw customer feedback has already been safely stored in the database. **Generate AI Analysis (LLM Chain)** Next comes the most important step in the workflow: **AI analysis**. The **Generate AI Analysis LLM Chain** node sends the feedback data to OpenAI. The prompt instructs the AI to analyze the feedback and return a structured response. The input sent to the AI includes: * Service type * Rating * Feedback message From these inputs, the AI generates several outputs, including: * AI sentiment * AI category * Urgency level * Feedback summary * Priority score * Confidence score * Root cause * Recovery direction * Recovery message This workflow uses the **GPT-4.1 mini** model, although other models can also be used depending on your needs. To ensure reliability, the workflow also uses a **structured output parser**. This guarantees that the AI always returns valid JSON, which is essential for automation workflows. **Store AI Analysis in Airtable** After the AI generates the analysis, the results are saved into the **AI Analysis table** using the **Create AI Analysis** node. At this stage, the system has successfully stored both: * The original feedback * The AI-generated insights This structured data allows the team to easily review and analyze customer feedback later. **Conditional Branching: Positive vs Negative Feedback** Once the AI analysis is completed, the workflow splits into two different branches depending on the result. **Positive Feedback Branch** If the feedback is positive and the rating is greater than three, the workflow executes the positive branch. Again, three actions are performed. First, an **appreciation email is sent to the customer** thanking them for the feedback. Second, the **operational team receives an email** containing the feedback and AI analysis. Third, the workflow posts a **Slack message in the positive channel** so the team can celebrate good customer feedback. This branching logic ensures that the team reacts differently depending on the nature of the feedback. ///insert screenshot for positive feedback email result ///insert screenshot for positive slack message result **Negative Feedback Branch** This branch runs when: * The AI sentiment is negative, or * The rating is less than or equal to three When this condition is met, three actions are triggered. First, the system sends an **email to the customer** containing a recovery message. Second, an **email notification is sent to the operational team**, including both the customer feedback and the AI-generated insights so the team can respond quickly. Third, a **Slack message is posted to the urgent channel** to alert the team immediately. ///insert screenshot for negative feedback email result ///insert screenshot for urgent slack message result * * * ### Summary In this project, we built a **complete AI-powered customer feedback analysis system** using **n8n, Jotform, Airtable, and OpenAI**. The automation collects feedback from customers, analyzes it using AI, stores the results in a structured database, and automatically notifies both customers and internal teams. The key benefits of this system include: * Automatically analyzing customer feedback with AI * Structuring insights such as sentiment, urgency, and root cause * Storing all feedback and analysis in Airtable * Automatically notifying teams through email and Slack * Helping teams prioritize and respond faster Instead of manually reviewing every feedback message, teams can rely on AI to quickly highlight the most important insights and ensure that no critical feedback is missed. If you want to build this automation yourself, you can download the full **n8n workflow JSON** from the video description and import it directly into your own n8n instance. Thanks for reading, and I'll see you in the next project.