Stop Using Email Rules (Do This Instead)

About this video

Kevin argues that traditional email rules, folders, and labels fail because email is unpredictable and context shifts, so you end up managing rules instead of your inbox — rules can automate steps but cannot decide what matters. In partnership with Make, he builds an AI agent instead. At a high level the agent does three things: it watches his Gmail inbox and decides whether an email needs his attention, labels and quiets messages that need no action, prepares a draft reply for messages that need action but are not urgent, and only sends a Slack notification when something truly cannot wait. Setup in Make takes a few minutes: create a scenario, add a module that watches for new Gmail messages, add an AI agent using OpenAI’s GPT 5.2, give it a short set of instructions (he links the exact instructions he used in the description), and pass in the sender, subject, and email body. He gives the agent knowledge — access to a simple FAQ in a Word document or TXT file so it can draft accurate replies to common questions — and three tools: applying Gmail labels, creating draft replies, and sending a Slack notification, each named and described so the agent decides when to use it rather than following hard-coded rules. He runs three real emails through it: an informational message is labeled no action with no draft or notification; a non-urgent actionable message is labeled action required and gets a draft but no interruption; and an urgent, progress-blocking message gets the label, a draft, and a Slack alert. He stresses transparency — Make shows the agent’s reasoning for each decision — and closes with a distinction: if a process is fixed and predictable, regular automation is faster and more reliable; if it needs thinking, judgment, messy inputs, or changing context, use an AI agent. The same pattern applies to support tickets, form submissions, sales leads, and internal requests.

Key takeaways

• Rules automate steps but cannot judge importance; an AI agent in Make makes judgment calls about whether email needs you, needs you now, or can wait.
• The agent uses three tools — Gmail labeling, draft-reply creation, and Slack notification — and only fires Slack for truly urgent, progress-blocking messages.
• Setup is: watch Gmail in a Make scenario, add an agent (he uses GPT 5.2) with short instructions, and pass in sender, subject, and body.
• Give the agent knowledge such as an FAQ Word or TXT file so its draft replies to common questions are accurate.
• Visibility into the agent’s reasoning for each decision is what makes it trustworthy; use plain automation for fixed, predictable processes and agents where judgment is involved.

Episode details: published February 11, 2026 · 8:13 long. Summary written by Self Growth Videos from the episode itself. Business results and any income figures mentioned are Kevin Stratvert’s own claims, shared for education — they are not typical or guaranteed.

← Previous: How to Use Granola AI (Tutorial for Beginners) · Next: QuickBooks Online 2026 Tutorial for Beginners (EVERYTHING You Need) →

All Kevin Stratvert videos · Kevin Stratvert profile · Kevin Stratvert on YouTube