Getting into AI Overview through Link Insertion: Independent AI Response Test #5

If you’re here, you probably don’t need an explanation of what AI Overview is. But let’s define it anyway and move on.

AI Overview is a generative block in Google that builds answers directly in the search results, using texts from open sources.

Simply put, it takes chunks from articles, combines them, and presents them as a ready-made answer. For SEO, this is both a problem and an opportunity. We’ll discuss the opportunities later. The problem is that AI Overview “steals” traffic from top sites: users get an answer without clicking, and the CTR of classic search results falls. According to Similarweb, by summer 2024, the impact of AIO on organic traffic became noticeable. And we, as SEO specialists, do a lot of work every day to push clients into Google’s TOP-3.

But let’s digress from AI Overview and talk about something else — link insertion.

Link insertion is a link-building tactic where we add mentions and links to already published materials. What makes this tactic cool? The article is already indexed, already has traffic and trust, and you simply “inject” your brand into ready-made content. By logic of use, this is close to parasitic SEO, where you place yourself on third-party domains that already have authority. The only difference is that link insertion is considered a more organic tactic.

And here an interesting question arises: what if we combine these two things? Actually, that’s what became our hypothesis for the experiment: if we insert a brand mention into an article that is already cited by AI Overview, then maybe the brand itself will appear in the answer too. The logic seems simple, but it’s worth checking in practice.

Hypothesis

Our hypothesis sounded as direct as possible: Can you get into AI Overview through link insertion in already-cited LLM articles?

The idea didn’t come out of nowhere. Look at how the model itself works: it doesn’t invent content, but collects fragments from materials that Google considers relevant. So if a new brand appears in the cited source, the model has no reason to ignore it.

We saw indirect confirmations of this before the test.

1. First, AIO often pulled in even minor mentions of companies if they were in cited materials.
2. Second, classic SEO experience with E-E-A-T shows that new mentions in trustworthy articles can quickly affect brand visibility.
3. Third, Google doesn’t yet have separate “filtering” mechanisms for such mentions in AIO, meaning any content in the source is a potential candidate in the results.

It turns out that link insertion could become a bridge between classical link building and new generative search. Our task was to verify whether this logic really works, or if it’s just an assumption. Because during our pool of AI response testing, some of our hypotheses were indeed disproven.

Methodology and Test Progress

Before choosing articles to work with, we look at all LLM results for selected keywords and identify articles that the answers cite.

Then comes standard link-building team work: we contact webmasters and find out the prices for placement in these articles, and only after that do we decide on sites and place mentions of the chosen company on them.

Actually, we planned to take 3 sites and place 3 links from articles that are in Google AIO on each. But we couldn’t reach agreement with some webmasters, so we ultimately settled on one from the auto parts niche.

For the test, we chose an article that was already cited in AI Overview. This is ideal: we knew for sure that the source was used by the LLM, and could track even small changes. We took a set of keywords and for each one looked at AI Overview and what it cites. The starting picture looked like this:

Next, we contacted the site owner and agreed on link insertion. Eventually it looked like this:

In the material, we added a mention of the brand “Avtoprostavka” with the anchor “specialists of the company Avtoprostavka”. After that, we reindexed the article so Google would “see” the update. Then came a waiting period.

The timeline looked like this:

• August 14 — added the mention to the article. 
• August 19 — nothing appeared in AI Overview yet. 
• After the next indexing — the brand appeared in the answer.

And here’s the first observation: the system doesn’t react instantly. On August 19, several days after making changes, the mention wasn’t yet in AI Overview. This means that even if an article is updated, the LLM doesn’t immediately pull new fragments into the answer. For us, this is an important signal: AIO has a certain lag in display, and we’ll have to account for this when we start scaling the tactic.

But the result turned out to be clear and demonstrative: AIO really does pull new mentions from already-cited articles. And soon we got this picture:

Results and Thoughts

So yes, after the next indexing of the article in AI Overview, the mention of “Avtoprostavka” appeared, we can confirm that the hypothesis is proven. Our brand really did appear in the generative search results thanks to just one link insertion in an article that was already a cited source.

This is a small victory for understanding how to manage AI Overview, but it shows the direction: we can influence what AI Overview shows to users.

What does this test give us? First, confirmation that link insertion in relevant articles affects not only classical SEO, but also Google’s generative search results.

Second, it’s worth understanding the limitations. This is still just one case, but the result needs to be confirmed in other niches and on other sites. We’ve already started doing similar things for clients and will be able to compile a complete picture soon of how successfully this works in various conditions. But even now it’s clear: the direction is promising, and it seriously changes the understanding of the “value” of link insertion.

Conclusions and Next Steps

What should businesses and marketers take away from this experiment? Sit down, open Google, and review your key queries and what’s currently in Overview. If you see an opportunity for your brand, contact the site owners and try to arrange link insertion. If you don’t have time for this, reach out to good link-building agencies or to us.

Our team is already testing new hypotheses and will publish soon:

• How dislike ratings of GPT responses affect its behavior; 
• Whether GPT can display content from a site that is closed to indexing in Google and Bing.

Author

Dmytro Kovshun

Dmytro Kovshun is the founder of Luxeo Team – an SEO Outsourcing Company. As a leading specialist in the industry, he is recognized as an expert in SEO promotion of websites. With years of experience and a deep understanding of the field, Dmytro continues to drive success and innovation in SEO strategies, helping businesses achieve their online goals.