The problem this solves
The newsletter is the marketing asset most often killed by its own cadence. Issue one is easy because enthusiasm writes it. Issue nine requires what every issue requires - hours of reading across the category, judgment about what mattered, and the writing itself - in a week when a launch or a fire drill claims those hours. So the most common newsletter in existence is the lapsed one, and a lapsed newsletter teaches the exact audience you most wanted to keep that your voice is optional.
The invisible cost is the research. Readers subscribe to newsletters that consistently know what happened and what it means, and that knowing is built from hours of scanning that produce three paragraphs. When time runs short, the research is what gets cut, and the issue degrades into links with adjectives. Readers detect the difference immediately even if they cannot name it, and the unsubscribe follows the third thin issue.
The durable division of labor keeps judgment human and makes the sweep automatic. An agent reads the week across your category, surfaces what genuinely moved, and writes the issue in your voice with a clearly argued take. You read it as editor rather than author, sharpen the take where your conviction differs, and send from your own address. The newsletter ships weekly because the expensive hours stopped being yours.
How the mission runs
- The week is swept, wide and then hard. Web Research covers your category's week: announcements, funding, product moves, notable analysis, and the conversations practitioners actually had. The sweep is deliberately wider than the issue, because knowing what to leave out is most of what makes a newsletter feel authoritative.
- Three insights survive the cut. From the sweep, the agent selects the three developments that matter for your specific readership and writes each as an insight: what happened, why it matters to them, stated in your voice. Selection is ruthless by instruction - a newsletter that includes everything informs no one.
- One take is argued, in your voice. The issue's centerpiece is a position: what this week's pattern means, argued with the evidence the research produced, sounding like you on a strong day. The take is drafted to be sharpened rather than to be safe, because the take is why readers open the next issue.
- The issue lands as a Gmail draft. The complete newsletter - subject line, three insights, the take, your standing sections - is written into a Gmail draft in your account. It sits with your other drafts, formatted and ready. The draft placement is the review gate: nothing sends until you do.
- Your edit is the editorial pass. You read as editor: sharpen the take where your view differs, cut an insight that misses, adjust a phrase that is almost your voice. Then you send from your own address, to your own list, on your own timing. Corrections you make consistently become instructions for future issues.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A publishable newsletter draft in your Gmail account each week: three researched insights from the category's actual week, one argued take in your voice, and a subject line, all formatted and ready. Your role compresses to an editorial read and a send. The research hours that made the newsletter unsustainable now happen without you, while every word that reaches readers still passes through your hands.
Make it yours
- Add a standing "who is hiring" or "tool of the week" section: the sweep already surfaces the raw material, and the agent fills the section each issue.
- Run a monthly deep edition alongside the weekly: one insight expanded into a full analysis, drafted from the accumulated month of research.
- Feed reader replies back in: forward the interesting responses, and the next issue's take can engage the sharpest disagreement by name or anonymously.
Frequently asked questions
Why a Gmail draft instead of sending it for me?
Because the send is where accountability lives. The draft-first design guarantees a human editorial pass on every issue: the voice check, the judgment call on the take, the final cut. An issue that ships without your fingerprints is a risk to the asset this playbook exists to protect.
How does it learn to write the take like me?
From your published writing and past issues, the agent builds a voice profile covering cadence, vocabulary, and how boldly you stake positions. Your weekly edits are the training signal that matters most: patterns in what you change become standing instructions, so drafts converge on needing lighter touches.
What if the category had a quiet week?
The agent says so honestly and adjusts: a quieter issue might carry two insights and a stronger take, or reach back to develop a slower-moving theme the sweep has been tracking. Manufacturing significance from a flat week is exactly the thin-issue failure the research pass exists to prevent.