How to Write AI Content That Ranks in Google & AI Search
Summary
SE Ranking published a video arguing that content produced with AI help can be indexed by Google and cited in AI search, and that the deciding factor is what happens after the draft. The team describes 16 months of testing in which pure AI pages were pushed to new domains and compared against AI-drafted articles that human editors reworked before publication on the company blog. In their account, 71% of the pure AI pages were indexed and picked up an early ranking window, then many of those pages fell away after roughly three months, while the edited articles held rankings and earned AI Overview citations.
The video turns that into a five part working method: get the technical and trust fundamentals right, write to a topic rather than a single keyword, structure pages so a model can lift a clean passage out of them, add the firsthand material a model cannot invent, and cite claims while linking internally with intent. The canonical link points at a GitHub repository holding a free Claude plugin and an SE Ranking MCP connection meant to run that method.
Demonstration by SE Ranking, embedded from YouTube. AIOTruth did not reproduce the test.
Why it matters
The practical consequence is that the question stops being whether a machine touched the draft and becomes whether the finished page carries something retrievable and attributable. A brand chasing citations gets no protection from disclosure or from avoiding AI entirely, and no penalty for using it: what determines whether a passage gets pulled into an AI answer is structure a model can quote cleanly, claims with sources attached, and material that exists nowhere else, such as your own measurements, your own customers, and your own failures. The reported pattern of early indexing followed by decay a few months later is the part worth planning around, because it means a page that ranks in its first weeks is not yet evidence that the approach works, and a content programme judged on that first window will keep producing pages that quietly disappear.
Source
- github.com/seranking/marketing-automation-plugins/tree/main/plugins/content-skills the item itself, published by SE Ranking, created by SE Ranking
- docs.github.com/en/site-policy/github-terms/github-terms-of-service Public source code or repository, first party
- github.com/features/copilot Public source code or repository, first party
- github.com/ikawrakow/ik_llama.cpp/pull/2369 Public source code or repository, first party
- github.com/aic0d3r/neon-ladder Public source code or repository, first party
- github.com/orgs/ultralytics/discussions/25250 Public source code or repository, first party
- github.com/ZC502/narh-yolo-align Public source code or repository, first party
- seranking.com/blog/ai-content-experiment/ Reputable reporting
- youtube.com/watch?v=lMnU9KIJSrcwhere AIOTruth found it
What was checked
- read the item page (200)
- https://github.com/seranking/marketing-automation-plugins/blob/main/plugins/content-skills/CHANGELOG.md: does not mention the item, tier 6 (Public source code or repository)
- https://docs.github.com/en/site-policy/github-terms/github-terms-of-service: supports the item, tier 6 (Public source code or repository)
- https://github.com/seranking/marketing-automation-plugins/tree/main/plugins/content-skills: does not mention the item, tier 6 (Public source code or repository)
- https://github.com/features/copilot: supports the item, tier 6 (Public source code or repository)
- https://github.com/features/ai/github-app: does not mention the item, tier 6 (Public source code or repository)
- https://seranking.com/blog/ai-content-experiment/: supports the item, tier 9 (Reputable reporting)
Limitations
- The demonstration was not independently reproduced by AIOTruth.
AIOTruth judgment
The item scores 8.5 with an evidence status of Supported. Usefulness at 10 reflects that the five moves are specific enough to apply the same afternoon and are attached to shipped tooling rather than left as advice. Relevance at 8 and evidence at 8 come from the publisher running the comparison on domains it controls and reporting a stated indexing rate, a decay window, ranking positions, and citation counts, which is more than most commentary on this question offers. Originality sits at 7 because the conclusion, that human editing and firsthand expertise are what separate performing pages from discarded ones, is now widely argued elsewhere, and the contribution here is the measurement rather than the idea. The reported figures come from the publisher's own blog and its own test domains, they were not independently reproduced by AIOTruth, and the sources checked alongside this item were GitHub platform pages rather than the underlying test data.
How this score was calculated
| Dimension | Weight | Score | What produced it |
|---|---|---|---|
| AIO relevance | 30% | 8 | ai search, ai mode; 5 scope question(s) matched |
| Usefulness | 30% | 10 | 3 artifact(s), 6 actionable marker(s), 2 measured figure(s) |
| Evidence | 25% | 8 | Public source code or repository; 3 verified source(s) across 2 domain(s) |
| Originality | 15% | 7 | carries original testing or data; closest archive match 0.1 |
aioRelevance x 0.30 + usefulness x 0.30 + evidence x 0.25 + originality x 0.15. The rubric is published in full on the Editorial Method page. Scoring is deterministic: the same item scores the same on every run.
Watch the demonstration Open the source
Found something wrong here? AIOTruth corrects material errors openly. Have something we should review? Submit a find.