How to run a weekly AI-assisted newsletter people actually open

TLDR: The newsletter that survives is not the best written one, it is the one that ships every week without burning you out. The system: Kagi to find topics and sources without noise, YesChat to draft with the model that best matches your register, Getimg.ai for a recognizable visual header, and a production template that turns it all into two hours of weekly work. Starting costs zero dollars.

This guide is for anyone using a newsletter as a marketing channel: the personal brand growing an audience, the business nurturing leads or the shop that wants back into the inbox every week. The problem with newsletters is not writing one: it is writing fifty-two.

AI solves the grind, not the judgment. A newsletter that is pure generated text reads like it and gets ignored. One that uses AI to research faster and draft the scaffolding, while the human supplies judgment and stories, keeps open rates high at a fraction of the effort.

1. Set up the topic capture system

Thursday’s blank page is avoided during the rest of the week. The system is a shared list where everything that could become a topic lands: customer questions, sector news, something about your market that annoyed you. AI enters at the research stage: when a topic looks promising, it needs verifying and enriching.

There, Kagi outperforms a free search engine, because newsletter research is professional search repeated weekly: its lenses filter by forums when you want what real people think, or by technical sources when you need the original data instead of the rehash. With no ads and no content farms, finding a story’s primary source takes minutes. From $5 a month with a free trial.

Always note the source next to the topic. A newsletter that links primary sources builds an authority no publishing volume can substitute.

2. Draft with AI without losing your voice

Before generating anything, write your style guide in ten lines: how you greet, how long your sections run, which words you would never use, how you sign off. That document gets pasted at the start of every drafting session, and it is what separates your newsletter from every other one generated with the same models.

YesChat brings its differentiating feature here: running the same brief through GPT and Claude and keeping whichever best imitates your register. Each model writes differently, and for text with a personal voice that choice shows. Generate section by section (not the whole issue at once) and edit the output by hand: cut what you would not say, add the anecdote only you have. The free plan covers the method, the paid one starts at $8 a month.

The golden rule: AI drafts the informational sections, you write the opening and the opinion. The two pieces that decide whether people read you next week are not delegated.

3. Produce the header image with Getimg.ai

The header image is brand, not decoration: the same visual line week after week makes your email recognizable before the subject line is read. With Getimg.ai you define a style prompt (palette, technique, mood) and each week you only swap in the issue’s topic, with 11+ models available to find the one that best executes your line. From $10 a month, with commercial rights on paid plans.

If your visual volume grows toward ads or social, the same prompt system recycles: the AI ad creatives guide uses the same base-prompt-and-variants logic. And to compare generators before choosing, there is the best AI image tools ranking.

4. Close the weekly cycle with a template and metrics

The full system fits in a production script: Monday, fifteen minutes to pick a topic from the list and verify sources with Kagi. Wednesday, one hour to generate sections with YesChat and edit them. Thursday, half an hour for the header, the subject line and the send. Total, about two sustainable hours a week for years.

Metrics close the cycle: opens to evaluate subject lines, clicks to evaluate topics. After four issues a pattern emerges of which topics move your audience, and that signal feeds back into the step 1 list. The newsletter improves on its own if the cycle holds.

This channel feeds and is fed by the rest of your marketing: AI competitor research generates topics your audience appreciates, and the AI for marketing and social media hub gathers every task in the sector. The rest of the productivity tools are in the best AI productivity tools ranking.

Common mistakes

Publishing generated text unedited. This is the terminal mistake: readers detect voiceless model prose and the newsletter moves to the ignored pile. Hand editing is not optional.

Changing the format every month. A newsletter is a habit product: same sections, same header, same day. Creativity goes into the content, not the structure.

Writing for everyone. The topic list degrades toward the generic without one concrete reader in mind. Return to the question of who you are writing to whenever a topic feels “generally interesting”.

Ignoring the click signal. Opens flatter and clicks inform: a topic nobody clicks does not interest them however often it gets opened. Adjusting the topic list with that signal is the cheapest improvement available.

Frequently asked questions

How much does this system cost to run?

Starting: nothing, since YesChat has a free plan and the header can wait. The full system with Kagi, paid YesChat and Getimg.ai lands at around $23 a month, less than a single piece commissioned from a copywriter.

Will people notice I use AI?

They notice unedited text, not the tool. If the opening and the opinion are yours and the generated draft passes through your edit, the result is your voice writing faster, and no reader penalizes that.

How often should I send?

The frequency you can sustain for a year. Weekly is the standard for a reason: enough presence to build habit, enough margin to protect quality. A fortnightly newsletter that ships beats a weekly one abandoned in March.

Which metric matters most at the start?

Clicks on topics, because they show what your audience wants. Open rates depend too much on subject lines and on each email provider’s inbox to guide editorial decisions early on.