Product Notes: MCP for product managers — connecting AI to Slack, Notion, and WhatsApp

Mohit Aggarwal published this article in the Product Notes publication on Medium as an introduction to the Model Context Protocol (MCP) aimed specifically at product managers rather than engineers. The piece is grounded in a concrete friction point: most AI-assisted PM workflows involve significant manual effort moving context between tools — writing something in Notion, copying it into Claude, editing the output, then pasting it back into Slack. MCP is presented as the infrastructure that removes this friction by giving AI systems direct read and write access to the tools a team already uses.

What the article is about

Aggarwal frames MCP not as a developer topic but as a productivity shift for non-technical professionals. He walks through the concept of protocol-based connections: instead of treating Claude or ChatGPT as isolated chat interfaces, MCP turns them into agents that can read from your Notion database, post to Slack channels, or query WhatsApp conversations based on a single instruction.

The practical examples he describes include having an AI agent generate a PRD from research already stored in Notion, update a project status without manual entry, or pull conversation threads from Slack before a planning meeting to surface open questions. The key point throughout is that none of this requires a PM to write code — the value comes from understanding what the protocol enables, then deciding how to structure work to take advantage of it.

Key takeaway

Product managers working with AI tools today spend a meaningful share of their time acting as translators between those tools and their actual work context. MCP shifts this: the AI system can hold context across tools persistently, rather than only within a single session window. This changes what kinds of questions PMs can reasonably ask AI assistants to help with during a workday, and how much setup each request requires.

Who it is useful for

This piece is for PMs and product leaders who are already using AI daily but feel the friction of disconnected tools. It does not require any technical background and is a direct starting point for understanding why MCP is worth paying attention to as it becomes more widely deployed across enterprise software.