Proof SEO is not AEO: Stitch AI hits Google with 9 views and an AI Overview citation

Page note — not spoken on film. Stitch AI (stitch-ai.com) is the client: a WhatsApp messaging product. Question.Marketing is the AEO / GEO agency on this engagement. Do not confuse the two brands. Hard receipts on film: Google Search shows the Stitch YouTube result with 9 views · 4 days ago; Google AI Overview cites stitch-ai.com and that YouTube video on WhatsApp PECR / bulk WhatsApp UK law queries. Soft / spoken-only (not title material): YouTube prominence in about twenty minutes; US friends seeing the same; industry conversion chatter. No "72 hours" claim on this film.

Proof that SEO is not AEO or GEO. I have been talking a lot about whether SEO is really viable anymore, because we are moving into an era where search engines are becoming redundant. We are very quickly at the point where answer engines (ChatGPT, Perplexity, Claude, Google AI Overviews, Gemini, and the rest) are becoming the norm.

What I wanted to demonstrate is part of an ongoing case study for one of my clients, Stitch AI. They are principally a WhatsApp messaging tool: bulk WhatsApp messaging, customer communication, and CRM plugs.

They have an article I published for them on their blog: bulk WhatsApp marketing UK law, with video, transcript, PECR consent, and TPS. We published that blog on the 15th of September. I also published a companion YouTube video on sending bulk WhatsApp marketing messages compliant with UK data protection and TPS rules. The answer, in short: yes, provided every person you message has specifically consented.

There was a gap in the knowledge base for AI where this content was not cited. So we published it to see what happened. I went live with Paris Childress of GEOforge (a product for finding where knowledge gaps exist) and I cannot highlight enough the need to add to knowledge bases if you have knowledge nobody else has.

Within this video (albeit an AI-generated video) there are about twenty questions answered around the topic. We are not trying to rank only for one horrendously long phrase. Yes, it is a long-tail question. The point is there is a knowledge gap: someone has asked this question before and there is not a reasonable answer AI feels it can rely on.

If you are in an industry where people are asking questions and the answers out there are rough, that is where you need to look. Tools such as GEOforge help find gaps. Can you add information gain: content that satisfies an answer more completely than what already exists? You cannot fill that gap by only regenerating what is already in the AI corpus. Google, ChatGPT, Claude, Perplexity, and Muse pull from different datasets. You would hope much of it comes from the open internet, but they are not identical libraries.

This is why I say classic search-engine optimisation, as the old game, does not matter the way it did. SEO, it is over, as a slogan. You still need good practice: the right content, consistency, and ideally third-party citations. One of the important things you can do for AI ranking is consistent data across multiple web properties that confirms what you say is true. You cannot publish AI slop, reword it, stick it on a page or a video, and hope it ranks. You need unique knowledge or a better answer to a question that has not been satisfied.

In our testing, if I publish a video only on YouTube it will not gain a citation in ChatGPT / OpenAI the way you might hope, because they are not using that the same way. You have to publish to multiple properties.

This piece went on the website about a week ago. The video is on YouTube, transcribed, with description, thumbnails, and images in the blog. Those images contain metadata and are not AI-generated. There is schema, Q&A, author data, confirmation across YouTube, LinkedIn, and other social properties. Google used to say social barely mattered; we are seeing that effect increase as AI search responses change.

I am going to prove it is not guessing. It is pulling information.

If I search whatsapp PECR regulations in the uk, that is not the headline of the video. If I take the long-tail bulk WhatsApp UK data-protection question and put that into Google, my result comes up: bulk WhatsApp marketing UK law. The video result is prominent on Google Search. That video was published about four days ago. It has only got nine views. You can see the Google result line: 9 views · 4 days ago.

What matters next: Stitch coming up in the AI Overview, with a citation. The video, linked to the web page, after about a week. I have never seen an SEO technique you can name that buys this cleanly. This is not cheating and not black hat. It is filling a knowledge gap.

If your SEO people tell you AEO and GEO are "just SEO", ask them what their strategy is and how they will attribute. We already hear that when products and services are cited in AI (Gemini, Google AI Overviews, Claude, Perplexity, Copilot, and so on), click-through traffic often converts several times better than ordinary web traffic. Figures people quote online vary by industry; treat ranges as industry chatter unless you have your own measurement.

Have a look at the fact that this video now comes up for different parts of the video: different phrases and questions answered inside it. On a broader query, the AI Overview cites the site again: blog and video recommended.

This Stitch project is only about six weeks in. We are building content, publishing it the right way, building external links, arranging podcasts, collaborating on social. There is a lot going on.

If you land on a piece of content and it answers your question perfectly, you are likely to deal with that company. If content is too broad, people bounce and you get a bad signal. Google wanted search to satisfy the need on the first result and never fully achieved it. AI can give a trusted answer and handle follow-ups.

Where we go next is interesting: answers will be tailored to the person. Search engines, as ten blue links, will matter less as Google pushes more people into AI chat. Showing how to become the answer is what fascinates me. This will change monthly, even weekly. Six weeks ago I would not have believed this was possible.

Hopefully that is useful. Please follow, please share, and I will see you in the next one.