The future of the inbox: Ogilvy’s rules for AI email (Martech World Forum 2026)
Manlio Carrelli’s Martech World Forum 2026 talk: AI now reads, summarizes and acts on email before people do. He carries David Ogilvy’s rules (respect the reader, be informative, sell, keep testing) into the AI inbox and adds a fifth: if you can’t trust it, you can’t ship it.
What is the talk about?
AI now sits between brands and their customers’ inboxes. Gmail and Apple Mail summarize email before anyone opens it, people buy straight from those previews, and agents read email and act on it for their owners. Manlio Carrelli, Stensul’s CEO, asks what won’t change, and answers with David Ogilvy, who called direct mail his “first love and secret weapon”.
What are Ogilvy’s rules for the AI inbox?
- The consumer isn’t a moron. Write for the reader and for the agent that reads on their behalf.
- Informative beats clever. Be specific, at scale: the bill is already a 1:1 email.
- We sell, or else. Email is owned and AI-native, so it is where AI-driven sales can be won.
- Never stop testing. Own performance after launch, not one subject-line test a quarter.
- If you can’t trust it, you can’t ship it. The new rule: build a system of trust.
What does a system of governance need?
AI can make thousands of email variants, but review capacity isn’t growing. Carrelli names four things a system of governance needs to close that gap: trusted ingredients built within approved guardrails, a fit with the stack where teams create and activate, rules enforced and risks flagged so human review scales, and results from each send that improve the next build.
Topics: AI & agents, Personalization, Brand governance
Make the move to agentic AI for email marketing.
We’ll show you how large enterprises are using Stensul to adapt their AI operating model, and ship more campaigns, with better results.
Request a demo

















































