Guides and explainers
Learn
How correspondence between assistants works — briefings, provenance, consent — and how to write one that lands.
Nothing matches. Clear the search to see every guide in this list.
Essays · 47
- How RelayLink keeps user inboxes separate
- Why RelayLink webhooks require HTTPS
- Why Webhook Payloads Should Not Carry Message Content
- Keeping the session off the record
- Why honest urgency gets answered
- The cost of an unlabeled assumption
- Making a follow-up a new ask
- When not to CC the whole team
- Why consent beats a shared inbox
- Writing so their assistant cannot invent your ask
- Closing a loop without another meeting
- When a meeting is still right
- Asking one question, not seven
- Separating facts from a guess
- Why the TL;DR is not the note
- What to do when you are the cold reader
- Writing for someone who reads on Monday
- Keeping private notes off the briefing
- Asking for a review, not a compliment
- When you need who said what
- Why a Slack thread loses the ask
- Why a shared doc is not a decision
- Handing context to whoever wasn't in the room
- Recording the option you rejected
- Stopping an assistant from sending as you
- When you use different assistants
- When the other person has no AI
- Asking without sounding like a bot
- Proving what you actually approved
- Getting a decision without scheduling a call
- Writing across companies with no shared Slack
- Why forwarding the chat fails at work
- When your two assistants disagree
- Respect the reader's time
- Urgency is a lie most messages tell
- Why correspondence needs a paper trail
- The reply is the point, not the send
- Your AI has opinions — not you
- The approval step is the product
- Email isn't broken — composition is
- AI-written email etiquette: what you actually owe the recipient
- Prompt injection is why your AI can't have a free send button
- Should you give your AI assistant access to your email?
- Can two AI assistants talk to each other?
- The most important words in AI correspondence: "verbatim, human-authored"
- AI-to-AI communication needs a consent layer, not a spam filter
- What is an AI briefing? (And why it beats forwarding a transcript)