---
title: "Stackby AI for Summarization & Categorization"
description: "Learn how to use Stackby AI for content summarization and auto categorization"
url: https://help.stackby.com/en/articles/450-stackby-ai-for-summarization-and-categorization
collection: "Stackby AI Field Agents - Use-cases"
updated: 2025-11-18
language: en
---

# Stackby AI for Summarization & Categorization

Learn how to use Stackby AI for content summarization and auto categorization

Stackby transforms raw data—long customer feedback, meeting transcripts, support tickets, and detailed notes—into **crisp, actionable takeaways** by seamlessly integrating advanced intelligence through AI Field Agents.

**Why Stackby’s Approach is Different: Bring Your Own AI**

Unlike most tools that lock you into a single, proprietary AI provider and pricing model, Stackby gives you complete control and flexibility:

-   **Bring Your Own Key:** Connect directly to leading, top-tier AI models from providers like OpenAI, Google Gemini, Anthropic.
    
-   **Full Control:** You decide how the AI is used, what specific prompts it runs (e.g., _“summarize into 5 bullets including action items and owners”_), and how the data is structured, ensuring outputs meet your exact business needs.
    
-   **Transparent Costs:** AI usage is billed directly to your chosen AI provider (via your own API key), not Stackby. This provides transparent, pay-as-you-go pricing without markups.
    

**How to Add AI for Summarization and Insight Extraction**

Here is the step-by-step process to set up an AI-powered data summarization workflow using Stackby's no-code environment.

Step 1: Set Up Your Data Repository

Start with a foundational, relational database structure within Stackby to organize your unstructured data.

1.  **Create a New Stack:** From your Stackby dashboard, start a new "Stack" (database).
    
2.  **Create Table:** Create a table to house your source data.
    
    -   **"Customer Feedback":** Store data like **Date, Customer Name, Sentiment** (Single Select), and a **Long Text** column for **Detailed Notes/Feedback.**
        
    -   _(Alternative: "Meeting Notes," "Support Tickets," etc.)_
        

Step 2: Add AI-Powered Columns (AI Field Agents)

Integrate intelligence directly into your data tables by adding the specialized AI column type.

1.  **Select Column Type:** In your data table (e.g., "Customer Feedback"), click the **\+ icon** to add a new column.
    
2.  **Choose "AI" Field Agent:** Select the **AI column type** from the column properties menu. This is Stackby's AI Field Agent, configured to perform the summarization task.
    
3.  **Name Your Column:** Give the column a clear, descriptive name reflecting the output, such as **"TL;DR Summary,"** **"Key Action Items,"** or **"Feedback Theme."**
    

Step 3: Connect Your AI Provider (Bring Your Own Key)

This is the control step where you link your Stackby workspace to your external AI service.

1.  **Find Your API Key:** Obtain your unique API key from your preferred AI provider (e.g., OpenAI, Gemini, Anthropic).
    
2.  **Paste Key into Stackby:** When you configure your new AI column, you will be prompted to paste your API key. Once linked, Stackby acts as the conduit, running your prompts against the provider, with all usage and billing managed by that provider.
    

Step 4: Craft Your Custom AI Prompt

The power of Stackby's AI lies in the flexibility of its prompting. You instruct the AI precisely on what to find and how to present it.

1.  **Define Your Prompt:** In the AI column configuration, write your instruction to the AI model.
    
2.  **Use Column Variables:** Reference existing column data (e.g., the long text field) directly in your prompt using brackets **\[ \]** to ensure context.
    

<table data-id="6301be84-eb9a-4112-a604-388cf3aac2ec" style="min-width: 75px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="1acd7d92-3e5b-4273-9bd5-de72586739ba"><span style="color: #1b1c1d;">Target Table</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="cba6f006-19ea-47d6-8126-21880a09b09d"><span style="color: #1b1c1d;">Output Column</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="fdf4d654-ab45-4e3f-b847-4884e9e25f6e"><span style="color: #1b1c1d;">Example Prompt</span></p></td></tr><tr><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="621b4fc4-f863-4c97-af5a-999192cfbeba"><span style="color: #1b1c1d;">Customer Feedback</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="6ac9079a-e341-44dd-a51c-2472483366d0"><span style="color: #1b1c1d;">TL;DR Summary</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="ec0ba99b-7b05-4d13-bbd7-5f6f136b7628"><span style="color: #1b1c1d;">"Read the [Detailed Notes/Feedback]. Provide a 2-sentence summary capturing the main problem and the desired outcome. Tone: Concise."</span></p></td></tr><tr><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="45168445-8894-43e3-a565-bd64453bb85f"><span style="color: #1b1c1d;">Meeting Notes</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="9696ec40-ed49-4d3c-956c-c09c363804fc"><span style="color: #1b1c1d;">Action Items</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="13d73837-cff7-4b43-8888-2be2e118807a"><span style="color: #1b1c1d;">"Based on the [Meeting Transcript], extract all explicit action items, including the person responsible. Return the list as a bulleted list of only 'Action: @Owner'."</span></p></td></tr><tr><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="eaa5cbad-3b44-4754-b56c-38b28be850ac"><span style="color: #1b1c1d;">Support Tickets</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="623ba39e-75a8-444b-91ab-f51e1b87d11a"><span style="color: #1b1c1d;">Priority Flag</span></p></td><td colspan="1" rowspan="1" style="vertical-align: top; background-color: #f8fafd; padding: 6pt 9pt; overflow: hidden; overflow-wrap: break-word; border-width: 0.6pt; border-style: solid; border-color: #1b1c1d;"><p data-id="876caf14-6228-4e7b-a0bf-9d2d608ad93d"><span style="color: #1b1c1d;">"Analyze [Detailed Notes/Feedback]. Classify the priority level as one word: 'Low,' 'Medium,' or 'High,' based on customer urgency and severity of the issue."</span></p></td></tr></tbody></table>

**Advanced: Incorporating AI Summaries into Automations**

Leverage Stackby Automations to run summarization jobs automatically and route the insights to the relevant team or dashboard.

-   **Weekly Digests:** Trigger a job every Monday to **summarize** all **"Meeting Notes"** from the previous week and send the output via a **Slack or Email notification** to leadership.
    
-   **Real-time Tagging:** On **New Record Created** in "Support Tickets," automatically run the AI Field Agent to extract the **"Priority Flag"** and set the ticket's **Status** based on that flag.
    

By using the AI Field Agent, you turn sprawling, qualitative data into structured, easy-to-read fields, transforming your Stackby workspace into an intelligent analysis powerhouse.