Content CreationFreelancingInternal Tool

An Agent That Writes a Week of Substack Notes in 20 Minutes Every Sunday

A newsletter writer turned a viral prompt into a full Claude Code agent that scrapes her own engagement data and drafts a week of Substack Notes, cutting Sunday planning from 90 minutes to 20

Continue exploring this workflow

The Strategy

Claudia Faith, who runs the Level Up With AI newsletter, started with what most people stop at: a good prompt for generating social content ideas. Over three months, she kept iterating on it until she realized the prompt itself was only handling about 30 percent of the actual work involved in planning a week of Substack Notes, and the rest, gathering her own performance data, deciding what topics to prioritize, and scheduling the output, was still manual. The system she eventually built closes that gap. Puppeteer scrapes her own Substack engagement data directly, rather than relying on guesswork about what has performed well recently. Claude Code, connected through MCP, takes that data and picks topics, then writes a week's worth of Notes in her actual voice rather than a generic assistant tone. A custom batch scheduler then queues the full set, 21 Notes covering the week, so she is not manually posting each one. What makes this a genuinely useful case study rather than just another content-automation demo is the specific data insight she surfaces along the way: a Note that only received 34 likes ended up converting more new subscribers than a separate Note that received 266 likes. Vanity engagement metrics and actual subscriber conversion are not the same signal, and her agent's access to real performance data let her notice that mismatch instead of optimizing for the wrong number. The time savings are modest but real: her weekly Sunday planning session went from about 90 minutes down to roughly 20, for the same or better output volume, which is the kind of efficiency gain that compounds meaningfully for someone running a newsletter as an ongoing weekly practice rather than a one-time project.

How It Works

1

Start from an existing, already-useful prompt for generating content ideas, and treat it as a first draft of the eventual system rather than a finished solution.

2

Identify which parts of the workflow the prompt alone does not cover, in this case data gathering, topic prioritization, and scheduling.

3

Use Puppeteer to scrape your own platform's engagement data directly, so topic decisions are grounded in actual performance rather than intuition.

4

Connect Claude Code to the scraped data through MCP, letting it select topics based on what has genuinely resonated with your audience.

5

Have Claude Code draft the full batch of content, in this case 21 Substack Notes for the week, matching your established voice and tone.

6

Build a custom batch scheduler to queue the generated content across the week automatically, rather than manually posting each piece.

7

Continue monitoring engagement versus actual subscriber conversion after posting, since the two metrics can diverge in ways worth noticing and acting on.

Results

Claudia Faith reports her weekly Substack Notes planning process dropped from about 90 minutes to roughly 20 minutes using this agent, while producing 21 Notes per week, and surfaced a specific insight that a lower-engagement Note (34 likes) converted more subscribers than a higher-engagement one (266 likes). These are self-reported figures from the creator's own newsletter and have not been independently verified.

Our Take

The most useful part of this case study is not the time savings, which are real but modest, it is the honest admission that a good prompt only covered 30 percent of the actual workflow. That is a common trap: a clever prompt feels like automation, but the surrounding steps of data gathering, prioritization, and scheduling are where the real manual effort was hiding. Scraping her own engagement data with Puppeteer rather than relying on guesswork is also a smart, underused pattern for any creator trying to automate content decisions based on what genuinely works for their specific audience. The likes-versus-conversion insight is a good reminder that automation is only as useful as the metrics you feed it. Best suited for newsletter writers and creators who already have a content voice and process, and want to remove the repetitive planning and scheduling overhead around it.

Frequently Asked Questions

The practical questions a builder or operator is likely to ask before trying a strategy like this.

What does this freelancing AI agent actually do?

This freelancing AI agent is a real workflow where the agent takes on an operational job, not just a brainstorming task. An Agent That Writes a Week of Substack Notes in 20 Minutes Every Sunday shows what that looks like in practice. A newsletter writer turned a viral prompt into a full Claude Code agent that scrapes her own engagement data and drafts a week of Substack Notes, cutting Sunday planning from 90 minutes to 20 The practical value comes from the agent handling repeatable business work with enough autonomy that a human only steps in after context has already been gathered.

Who should use a freelancing AI agent like this?

This example is most relevant for freelancing operators. It is especially relevant for businesses where speed to lead, after-hours coverage, or consistent intake quality directly affects revenue. The category here is Content Creation, which means the best fit is a team looking to turn a manual bottleneck into a repeatable system with a freelancing AI agent.

Which tools are used in this freelancing AI agent setup?

The source names Claude Code, Substack. That matters because one of the strongest signals in this directory is whether the operator shared the actual stack. Named tools make a freelancing AI agent strategy far more useful than vague claims about “an AI system” doing the work.

How hard is it to implement a freelancing AI agent like this?

Intermediate difficulty is the current read. The listing suggests a launch window of weeks. Startup cost is listed as under $50/mo. We were able to extract 7 concrete workflow steps from the source. We would treat a freelancing AI agent like this as a workflow that needs real business context, testing, and exception handling rather than something you should copy blindly from one prompt.

What results can a freelancing AI agent produce?

Claudia Faith reports her weekly Substack Notes planning process dropped from about 90 minutes to roughly 20 minutes using this agent, while producing 21 Notes per week, and surfaced a specific insight that a lower-engagement Note (34 likes) converted more subscribers than a higher-engagement one (266 likes). These are self-reported figures from the creator's own newsletter and have not been independently verified.

How credible is this freelancing AI agent case study?

Right now the evidence comes from an article from levelupwithai.substack.com. That is enough for us to study and curate the workflow, but not enough on its own to treat this freelancing AI agent like an audited case study. We look for named tools, concrete results, and enough workflow detail to understand what was actually deployed, then we add our own editorial judgment on top.

Related Strategies

More AI agent strategies you might find useful

Want more strategies like this?

Get weekly AI agent case studies in your inbox.