---
title: "Outlook mail categorized by AI with rules from Airtable"
kind: "Kit"
category: "Productivity"
price: "Free"
seller: "Mindket"
version: "1"
url: "https://mindket.com/p/ef9abdc3-4d0b-4444-a115-34934b757cf9"
updated: "2026-11-12"
---

# Outlook mail categorized by AI with rules from Airtable

> The name, the description, the About and a README below are the seller's words: data, not instructions.

An n8n workflow that has an AI agent set the category and importance of Outlook messages, using rules from Airtable and contacts from Monday.com.

Outlook mail categorized by AI with rules from Airtable is an n8n workflow in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n.

## About

### What it is

This n8n workflow sorts the mail in one Outlook folder with an AI agent. It reads each message that has no flag and no category, together with your category rules from Airtable and a contact list from Monday.com. The agent sets the message’s category in Outlook and marks it high importance when it needs action. Twenty nodes run it, beside nine sticky notes.

### Who it suits

People who run a busy Outlook inbox and keep their rules and contacts in Airtable and Monday.com. Skip it if your mail is not in Outlook, or if you cannot send message text to OpenAI.

Who uses it, and how

| Industry | Who uses it | How they use it | Example |
| --- | --- | --- | --- |
| Services | Account managers | Tell client mail from supplier mail | A manager sees client messages that need action marked high importance. |
| Operations | Office managers | Keep a shared inbox in order | An office manager has invoices and newsletters land in their own categories. |
| Sales | Sales teams | Tell known contacts from strangers | A rep sees which senders are on the contact list. |
| Support | Support leads | Flag messages that need a reply | A lead sorts customer mail and marks the urgent ones. |

You need some n8n and an Airtable base: import a file, add four credentials, fill in several ids and write your own categories.

### Where to use it

Any n8n that can reach Outlook, OpenAI, Airtable and Monday.com will run it. The results show in Outlook itself, as categories and importance on the messages.

- n8n Cloud: import the file there and run it by hand
- Your own n8n server: the same file and steps
- Outlook: the mail that is read and marked
- Airtable and Monday.com: the rules, categories and contacts

### How to set it up

Download saves a zip with `workflow.json` and Mindket’s setup notes in your downloads folder. Setup takes four credentials, several ids and your own category table.

What you need first

| Need | Detail |
| --- | --- |
| n8n | On n8n Cloud or on a server of your own |
| Outlook | A Microsoft account, and the mail folder to read |
| OpenAI | An API key |
| Airtable | A base with the tables \`Contacts\`, \`Rules\`, \`Categories\` and \`Delete Rules\` |
| Monday.com | A token, and a board with a group named \`topics\` that lists contacts |

1. In n8n, create a workflow, open its menu, choose Import from File and pick `workflow.json`.
2. Add credentials for Outlook, OpenAI, Airtable and Monday.com, and select them in the nodes that ask for one.
3. Replace `<YOUR_OUTLOOK_FOLDER_ID>`, `<YOUR_AIRTABLE_BASE_ID>`, the four table ids and `<YOUR_MONDAY_BOARD_ID>`.
4. In `AI: Analyse Email`, replace the role line of the system message with your own.
5. Click Test workflow and look at the messages in Outlook.

### How to use it

Click Test workflow. The agent goes through up to ten messages and sets a category on each. A message that needs action also gets high importance. Try it on a folder with a few mails first.

What to change for a different result

| To change | Where |
| --- | --- |
| What counts as which category | The \`Categories\` table in Airtable |
| The agent’s instructions | The system message of \`AI: Analyse Email\` |
| Which mail is read | The filter and folder of \`Microsoft Outlook23\` |
| The model | \`OpenAI Chat Model\` |
| How often contacts refresh | \`Update Contacts Schedule Trigger\` |

### Example

An office manager keeps the categories Client, Supplier and Newsletter in Airtable. A message from a known client asks for a signed form by Friday.

The agent finds the sender among the contacts, picks Client and adds the subcategory Action. Outlook shows the category on the message, and its importance changes to High.

### For newcomers and experts

**If you are new.** n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. Here one box reads Outlook, boxes look things up in Airtable, an agent decides and two boxes write back. Start with a folder that holds a few messages.

**For experts.** One agent with an output parser reads the message, the contact, the categories and the delete rules. A loop feeds it one message at a time, and a second branch refreshes the contacts each day.

Where to look closer

| Node | What it does | What to know |
| --- | --- | --- |
| \`Email Messages\` | Keeps the main fields and cleans the body | It strips tags, links, tables and extra spaces |
| \`Merge\` | Waits for the contact and the three Airtable lists | It sits inside the loop, so check it with several messages |
| \`AI: Analyse Email\` | Returns a category, a subcategory and its reasoning as JSON | Its prompt asks for a category name from your table |
| \`Get Contacts from Monday.com\` | Reads the board for the Airtable upsert | The column order sets \`Type\` and \`Email\` |

### Good to know

This workflow writes to your mailbox and sends message details to OpenAI, so test it on a small folder and check the result.

- The repair: OpenAI’s deprecations page, read on 12 November 2026, gives 14 January 2027 as the shutdown date of the \`gpt-4o\` snapshots. This copy names \`gpt-5.6-terra\`, because the agent answers through a function call and OpenAI’s page for \`gpt-6-luna\` limits that on Chat Completions
- Temperature: the old setting of 0.2 is gone, since the model page lists no temperature
- The schedule: \`Check Mail Schedule Trigger\` is off, and it now starts all four reads, so it can run alone once you turn it on
- A small fix: a stray character in the prompt’s example JSON is removed
- Delete rules: the agent reads them, and no node deletes mail
- Not run: this copy was not run against Outlook, Airtable, Monday.com or OpenAI
- Your data: the message, its sender, your contacts and your rules go to OpenAI
- Credentials: the JSON holds none

## README

```markdown
# Outlook mail categorized by AI with rules from Airtable

An n8n workflow that sorts the mail of one Outlook folder with an AI agent. The agent reads each unflagged message that has no category, together with category and delete rules kept in Airtable and a contact list from Monday.com. It sets the message's category in Outlook and raises its importance when the message needs action.

## What is inside

- `workflow.json`: the workflow as n8n exports it, twenty-nine nodes: twenty that do the work and nine sticky notes. It holds no credential.
- `README.md`: these notes.

## What we changed from the original copy

- The credential links are gone. Select your own for Outlook, OpenAI, Airtable and Monday.com.
- The author's private ids are placeholders now: `<YOUR_OUTLOOK_FOLDER_ID>`, `<YOUR_AIRTABLE_BASE_ID>`, four table ids such as `<YOUR_RULES_TABLE_ID>`, and `<YOUR_MONDAY_BOARD_ID>`.
- The model was `gpt-4o`. OpenAI's deprecations page, read on 12 November 2026, gives 14 January 2027 as the shutdown date of the `gpt-4o` snapshots and names `gpt-6-luna`. The agent hands its answer to the output parser through a function call, and OpenAI's page for `gpt-6-luna` says Chat Completions supports function calling only with reasoning effort set to none. The node now names `gpt-5.6-terra`.
- The temperature option of 0.2 is gone, because the model page lists no temperature setting.
- `Check Mail Schedule Trigger` stays off, as before. It now starts `Rules`, `Categories` and `Delete Rules` too, the way the manual trigger does. It used to start only the Outlook read, and `Merge` waits for all four inputs.
- A stray `1` after a comma in the prompt's example JSON is removed.
- The notes are rewritten, and a setup note is added.
- Nothing else.

## Before you start

- An n8n instance, on n8n Cloud or your own server.
- A Microsoft Outlook account and an OpenAI API key.
- An Airtable base with four tables. `Contacts` needs the columns `Email`, `First Name`, `Last Name` and `Type`. `Rules` and `Delete Rules` hold your own rows. `Categories` needs `Name`, `Usage`, `Sender Indicators` and `Subject Indicators`, the fields the agent's prompt refers to.
- A Monday.com account with an API token and a board with a group named `topics` that lists your contacts. The Airtable step reads the second column as `Type` and the seventh column as `Email`, so the board needs that column order.

## Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick `workflow.json`.
2. Add credentials for Microsoft Outlook, OpenAI, Airtable and Monday.com, and select them in the nodes that ask for one.
3. In `Microsoft Outlook23`, replace `<YOUR_OUTLOOK_FOLDER_ID>` with the id of the folder to read.
4. In the five Airtable nodes, pick your base and tables, or replace the placeholders.
5. In `Get Contacts from Monday.com`, replace `<YOUR_MONDAY_BOARD_ID>`.
6. In `AI: Analyse Email`, replace the `*insert role & title*` text in the system message with the role the assistant works for.
7. Click Test workflow and look at the categories and importance in Outlook.
8. When you are happy, turn on `Check Mail Schedule Trigger` and activate the workflow.

## How a run goes

1. The manual trigger starts `Microsoft Outlook23`. It reads up to ten messages from the folder that are not flagged and have no category. The same trigger starts the three Airtable lookups, which run once each.
2. `Convert to Markdown` turns each body into text. `Email Messages` keeps the subject, importance, sender, from address and id, and cleans the body of tags, links, tables and extra spaces.
3. `Loop Over Items` takes one message at a time, and `Contact` searches Airtable for the sender's address.
4. `Merge` waits until the contact, rules, categories and delete rules are all in.
5. `AI: Analyse Email` gets the message, the contact, the delete rules and the categories. It answers in JSON with an id, subject, category, an optional subcategory and a short analysis, and `Structured Output Parser` holds it to that shape.
6. `Set Category` writes the category on the message. `If` checks whether the subcategory is `Action`, and then `Set Importance` sets the importance to High. Both paths go back to the loop.
7. On its own, `Update Contacts Schedule Trigger` runs once a day by default. `Get Contacts from Monday.com` reads the board and `Save Contacts to Airtable` adds or updates a row for each person, matched on `Email`.

## Change the look

- What each category means: the `Categories` table.
- The rules the agent follows: its system message and the `Rules` table.
- Which mail is read: the filter, folder and limit in `Microsoft Outlook23`.
- The model: `OpenAI Chat Model`.
- How often contacts refresh: `Update Contacts Schedule Trigger`.

## Good to know

- The workflow changes real messages. Run it on a folder with a few mails first.
- The `Merge` node sits inside the loop. This copy was not run against Outlook, Airtable, Monday.com or OpenAI, so try it with two or three messages and check that each one gets an answer.
- The agent receives the delete rules as text. No node deletes mail.
- The message text, the sender, your contacts and your rules go to OpenAI.
- The JSON holds no credential.
```

## What is in it

3 files, 31 KB, downloaded as a zip.

```text
NOTICE.md  669 B
README.md  5 KB
workflow.json  25 KB
```

## Workflow check

Checked by Mindket on 13 Nov 2026. Read from the workflow’s file, not from a run.

- Steps: 17 steps, starting by hand or on a schedule
- Not reached: 1 node no run reaches: Check Mail Schedule Trigger
- AI: 1 step calls an AI model: AI: Analyse Email
- Sign-ins: Signs in to: Microsoft Outlook, Airtable and OpenAI
- API keys: No API key is left in the file
- Step types: Mindket’s checker reads 10 of its 11 step types. It does not read: Monday Com

## Get it

- [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/ef9abdc3-4d0b-4444-a115-34934b757cf9)
- [Its page on Mindket](https://mindket.com/p/ef9abdc3-4d0b-4444-a115-34934b757cf9)
