# Notion CRM Intake with Jev and Composio for AI Lead Management in 2026

Build a Notion CRM intake with Jev and Composio: classify, score, and route leads without an LLM. Python Colab notebook included.

Source: https://zakaz-04481.shop/notion-crm-intake/ · based on the channel [Agentic AI Hub](https://www.youtube.com/channel/UCFHkh7K7E39HejS_pV8NEfA) · Video: [How I Let AI Manage My Leads (Without an LLM) | Notion CRM Intake with Jev + Composio](https://www.youtube.com/watch?v=nY9RXUV2hxI) · 2026-11-13

![Notion CRM Intake with Jev and Composio for AI Lead Management in 2026](https://zakaz-04481.shop/notion-crm-intake/notion-crm-intake.webp)

## Key takeaways

- Jev by TypeSafe System One classifies, scores (0-100), and routes each inbound item as lead or not, with type and hot/warm/cold label plus confidence.
- Python extracts contact fields; Composio connects to Gmail, Notion, and Slack to perform actions.
- DRY_RUN is on by default: the notebook shows a ranked table and planned writes without creating a Notion database or posting to Slack.
- The pipeline takes inbound email or form entry, scores it, writes a Notion row, and pings Slack for hot leads.
- The Colab notebook is available at colab.research.google.com/drive/1nJepqWV0GcAIAaacFDIT8yMe9I1HKJTF.

You can set up a Notion CRM intake with Jev and Composio by running a Python Colab notebook that uses Jev (TypeSafe System One) to classify and score each inbound lead, then Composio to write the result to Notion and alert Slack for hot leads—all without a large language model. The notebook is available at [this Colab link](https://colab.research.google.com/drive/1nJepqWV0GcAIAaacFDIT8yMe9I1HKJTF), and DRY_RUN is on by default so you can see the ranked table and planned writes before any data is created.

## How the Notion CRM intake pipeline works from inbox to Slack

The pipeline has four stages: ingest, classify, score, and route. First, Python reads an inbound email or form entry and extracts contact fields such as name, email, company, and message. Then Jev classifies the item as a lead or not, assigns a type (for example, demo request, partnership, or support), and gives a 0–100 score with a hot, warm, or cold label, each with a confidence score. Finally, Composio uses that output to create or update a row in a Notion database and, if the lead is hot, posts a message to Slack. The entire flow runs in a single Colab notebook, and you can test it without side effects because DRY_RUN is enabled by default.

## What Jev is and what it cannot do

Jev is a classification and routing service from TypeSafe System One that returns structured labels and scores for each inbound item. It does not generate text: it cannot write replies, summaries, or Notion or Slack copy. Its only job is to decide whether an item is a lead, what type it is, how hot it is on a 0–100 scale, and how confident it is in those decisions. That narrow scope is what makes it reliable and fast for lead intake, and it means you don't need an LLM to handle the routing half of a CRM.

## Setting up the notebook with two keys and connections

To run the notebook, you need two API keys: one for Jev and one for Composio. Composio then handles the OAuth connections to Gmail, Notion, and Slack, so you don't have to manage tokens manually. The notebook walks you through entering the keys and authorizing each service. Once connected, you can run the dry-run cell to see a ranked table of leads and the planned writes to Notion and Slack. No database is created and no messages are sent until you turn off DRY_RUN.

## Step-by-step: running a real dry run and going live

1. Open the Colab notebook at the link above.
2. Enter your Jev and Composio API keys in the designated cells.
3. Authorize Composio to access Gmail, Notion, and Slack.
4. Run the dry-run cell: it processes sample inbound items and displays a ranked table with scores and planned actions.
5. Review the output to confirm classification and routing logic.
6. When ready, set DRY_RUN to False and run the live cell to create Notion rows and send Slack alerts for hot leads.

## Why lead intake does not need an LLM

Lead intake is fundamentally a classification and routing task: decide if an item is a lead, what type, how hot, and where it should go. These are structured decisions that a specialized model like Jev can make faster and more cheaply than a general-purpose LLM. An LLM is optimized for generating text, but here you need consistent labels and scores, not prose. By using Jev for the routing half, you avoid the cost, latency, and unpredictability of an LLM while still automating the intake process.

## Links and resources

- [Colab notebook (LeadScout)](https://colab.research.google.com/drive/1nJepqWV0GcAIAaacFDIT8yMe9I1HKJTF)
- [Free book: A-Z of Agentic AI](https://535214083743.gumroad.com/l/qnjoqo)
- [Agentic AI Hub YouTube channel](https://www.youtube.com/@AgenticAIHub1)

## Conclusion

The Notion CRM intake with Jev and Composio offers a practical way to automate lead management without an LLM. By combining Jev's classification and scoring with Composio's integrations, you can route leads to Notion and Slack in minutes. Start with the dry run to see the ranked table, then flip DRY_RUN to go live. For a full walkthrough, watch the video by Agentic AI Hub. Ready to build? Open the [Colab notebook](https://colab.research.google.com/drive/1nJepqWV0GcAIAaacFDIT8yMe9I1HKJTF) and run your first dry run today.

## Questions & answers

**What is Jev and how does it differ from an LLM?**

Jev is a classification and routing service from TypeSafe System One. It returns structured labels and scores for each inbound item, such as lead or not, type, 0–100 score, and hot/warm/cold with confidence. Unlike an LLM, it does not generate text; it only classifies and routes, making it faster and more predictable for lead intake.

**Do I need an LLM to manage leads in Notion?**

No. Lead intake is primarily a classification and routing task, which Jev handles without an LLM. You can extract contact fields with Python and use Composio to write to Notion and Slack, avoiding the cost and complexity of an LLM.

**What is DRY_RUN and why is it on by default?**

DRY_RUN is a mode that shows what the notebook would do without actually creating a Notion database, inserting or updating rows, or posting to Slack. It is on by default so you can review the ranked table and planned writes before making any changes.

**Which services does Composio connect to in this setup?**

Composio connects to Gmail for reading inbound emails, Notion for creating and updating CRM rows, and Slack for sending alerts about hot leads. It handles the authentication and API calls for these services.
