How I Do Content Engineering with Claude Code

Here’s how I built a high-quality content automation system for the Ahrefs blog using Claude Code and 23 skill files.

Back in August 2025, I shared the AI content process I had developed for the Ahrefs blog. It used ChatGPT projects and custom GPTs to speed up certain types of content creation from several days to a couple of hours, but still required tons of manual intervention.

Now, barely eight months later, I’m sharing our new process. I use Claude Code and 23 custom skill files, chained together, to generate publish-ready article drafts in six to twelve minutes. We have published around 15 articles with this new process, and updated some 30 or so more.

I’ve been using AI to help create content marketing since 2020. It has been useful in effortful, piecemeal ways. But today it is good enough to automate important parts of content marketing with no loss in quality (and even a significant gain in some areas, like research). Or as I put in a recent article: AI content wasn’t good enough. Now it is.

As a result, I suggested a pretty bold direction in our company Slack, back in February:

Here’s our AI current content process.

Watch this process on YouTube

Check out this episode of the Ahrefs podcast to watch me demo our content automation system to Ahrefs’ CMO, Tim Soulo.

LinkedIn comment put it, very articulately:

“Ryan’s SKILL files are good because Ryan already knew what to put in them. Most people using blank-slate tools don’t have 13 years of editorial experience to build from. The gap isn’t just in the tool. It’s in the person behind it too.”

Topic selection still matters

This process is geared specifically towards informational SEO content. I only use this process on topics that I understand well, so that I can review each article to validate its claims, correct misinformation, and make sure I feel happy putting it out into the world.

I also focus primarily on topics that Ahrefs has already covered (in some capacity), allowing us to use hundreds of existing, high-quality articles as a reference point for new content.

We have no plans to “scale content” with AI

I could use this process to scale the Ahrefs blog to tens of thousands of articles. I will not. It would not be in the interests of Ahrefs or our customers.

Instead, I am using this workflow to help us maintain an evergreen library of useful content on a handful of core topics. My goal is to remove drudgery and focus human grey matter on the parts of marketing that benefit most from it.

Ahrefs MCP, review and prioritize the best keywords to target, and kick off the entire blog-pipeline workflow, notifying me when new article drafts are ready for review.

The skill tests each stage of the process, generating outlines, research primers and drafts both with and without the guidance provided in our custom skill files.

The LLM reviews the outputs and makes suggestions for how to improve the skill file for more consistent results.

It’s easy for skill files to become long and bloated, and in doing so, make it less likely that their guidance will be correctly applied by the LLM. This process allows me to continually shorten skills to their most effective essence, and remove skills that don’t have any real bearing on my desired output.