When should you use it?
- You publish a regular newsletter and dread the grind
- You are not technical and want a fast setup
- You are fine reviewing drafts before they send
Not for: Publishing drafts without reading them; Topics needing deep original reporting the agent cannot do
What do you need, and what do you get?
You need
- A no-code agent tool and a schedule
- Access to research, a database and email
- A writing-style guide for your voice
You get
- A researched draft of the week's newsletter
- Repurposed posts for other channels
- Everything filed in a database automatically
How does it work, step by step?
- Set the triggerPick a schedule, such as every Monday morning.
- Write the instruction promptTell the agent to research your niche, draft content and store and email it.
- Grant the skillsGive it research, a database, email and a writing module to choose from.
- Add your voiceInclude a style guide so the drafts sound like you.
- Review before it shipsRead each draft; the agent decides tools, but you decide what sends.
The prompt
Our own prompt, inspired by Annika Helendi and written to work in Custom agent. Values in lime are placeholders.
Create a scheduled agent that runs every Monday at 9am. Its job: research this week's developments in [my niche], draft a newsletter for [my audience], and generate a few repurposed posts for LinkedIn and X. File the results in [database] and send the draft to my inbox. Use my writing style guide: [paste]. Choose whichever research and writing tools fit each step, and do not send anything to subscribers, only to me.
Inspired by Annika Helendi: Build your own AI newsletter AGENT in minutes. The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.
What can go wrong?
- The agent chooses tools, so vague instructions give vague output; be specific about the task.
- It drafts, it does not fact-check; verify claims before publishing.
Our take
Annika Helendi's main point is the contrast with traditional automation. Instead of mapping every branch in advance, you describe the goal and the agent chooses which research and writing tools to call, which is why setup took minutes rather than an afternoon of wiring. That flexibility is also where quality can slip, because an agent given a loose brief produces loose work. Write the instruction as you would brief a freelancer: audience, length, sections, sources you trust and sources to avoid. A short style guide with examples of your own past newsletters does more for voice than adjectives like friendly or professional. Ask it to list links for every claim so you can check them quickly, since research agents sometimes report trends from stale or thin sources. The Monday email to yourself is the right control; keep it as a draft for review, and never let the agent send to subscribers directly. Swap in your own industry for the topic and keep the database, because past issues become a useful archive for spotting repeated themes.
Where the time goes back: A weekly content grind that once ate a morning of setup and writing becomes a Monday draft you only need to review.
FAQ
How is this different from a normal automation?
Do I need to code?
Source and credit
Inspired by Annika Helendi. We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.
- Inspired by
- Build your own AI newsletter AGENT in minutes
- Creator
- Annika Helendi
- Published on
- annikahelendi.substack.com (Substack)
- Original date
- Jun 8, 2025
- License
- Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
- Originally for
- Custom agent
- AI dry run
- Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. PASS: Gates held on desk check; no live sends or purchases.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails. - Verified
- Oct 23, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.
Are you the creator? Claim or correct this page.