SEO on autopilot: How individual articles become a network of topics

Generating a hundred articles is no great feat these days. Anyone can do it in an afternoon for a few euros, and because anyone can do it, it’s hardly worth anything anymore.

The real work lies in the questions that come before: What articles does your community actually need, how are they connected, and which ones are best left unwritten? Aivor® Content now offers SEO topic clusters to address precisely these questions.

One overview article, surrounded by eight to fifteen detailed articles, all internally linked and planned based on an analysis of your existing website. The reason for this is hardly romantic. In most niches, a single article simply no longer ranks, no matter how well it’s written.

Two articles, one search query, both lose out

The real problem with previous content projects wasn’t the quality of the text, but a failure to recognise what already existed. Anyone who’s been running a forum for four years will already have content on their most important topics. If a generated article on the same keyword is then added, two of your own pages compete with each other, and Google usually ends up choosing neither.

That’s why Aivor first reads the sitemap and the existing articles before any planning takes place. If an existing article already covers the topic, a second one isn’t created: the existing one is expanded into a complete version and replaced in the CMS. The title and URL remain unchanged, because WoltLab generates the article URL from the title, and a new title would alter the URL of a page that’s already ranking.

The second mechanism comes into play within the plan. If two planned articles effectively cover the same Google results page, a warning appears for both topics, along with a button to delete them. This saves twenty credits.

The plan comes from Google, not from the language model

The topic research analyses real Google search queries, as well as the autocomplete suggestions in the search bar and related questions. If you connect your Search Console (which is free), you’ll also receive your own figures for the last 90 days: where the website is already appearing in the results and where it’s just missing out on page one. It is precisely in these areas that Aivor recommends expanding an existing article rather than writing a new one.

From this, the planner identifies three to five topic areas, one (or all) of which you can expand upon. The final plan then shows which of your keywords will get their own article, which will be included as a section, and which will be omitted. You can remove items right up until the time of writing, even after confirmation.

The most important feature of this update is a change of direction

Since 22 August, every full plan has included at least one news article and one science article. Previously, the planner suggested almost exclusively how-to articles, which was convenient but also rather monotonous. News articles are based on real news stories from the past seven days, whilst science articles are based on up to six studies from OpenAlex and PubMed, each with a bibliography.

More interesting is the scenario where the search turns up nothing. In that case, no article is produced. The topic is marked as incomplete, the ten credits are automatically refunded, and the system moves on to the next one. I consider this almost the most important change in the entire update, because a tool that would rather deliver nothing than something made up is the only one whose promise of ‘autopilot’ can actually be trusted.

The overview article continues to write itself

The plugin retrieves the finished articles hourly and then reports back to which address they have been published. This is what makes the internal linking work: new articles link to relevant existing ones, existing ones are subsequently linked to the new ones, and the overview in the pillar article updates itself automatically. For each existing article, Aivor places no more than one link every 30 days, so that maintenance doesn’t turn into link spam.

Aivor never accesses your community. Your server fetches the content.

And what does Google say about this?

It’s a valid question, especially when it comes to automatically generated content. Since 2024, Google has been taking action under the heading Scaled Content Abuse against large volumes of pages that exist primarily for the sake of rankings and offer no value to readers. How a page came into being is explicitly irrelevant in this context. The only thing that matters is whether it has intrinsic value.

The cluster approach is designed with this in mind. Each article serves its own search intent and stands apart from the rest; topics lacking substance are avoided from the outset, and the final plan passes through your hands before production begins.

Conclusion

Texts have become mass-produced commodities, but structure has not. Whether your community is discovered or simply has a lot of content depends on whether the articles work together.

The Aivor plugin: Automatic content import connects your WoltLab Suite with Aivor. As the initial website analysis is free, you can view the full topic plan before making a decision.

You can find details of the full range of features on the Aivor Content product page. You can find the complete step-by-step guide in the Aivor Wiki.