Ask yourself a harder question than “what should we publish this month?” Ask what you could publish that a model does not already know.
That question has an unhelpfully short answer for most content plans. If ChatGPT can write your post from its training data, it has no reason to go and read yours. Mike Montague put the mechanism plainly when he was on the show:
“if AI can write a blog for you, it already has that knowledge. So What why would it index it? Why would it cite your version if it can already write one that doesn’t need to be cited? It already has this domain knowledge. So what could we do that could add to AI, tell it something that it doesn’t know?”
— Mike Montague, Avenue 9, on Unscripted SEO
Conversations are the one input that has not been scraped
A transcript of a real conversation contains things that exist nowhere else: the number a practitioner actually saw, the client they actually fired, the tactic that worked and then stopped working. Mike calls the alternative approach degenerative engineering, and it is the opposite of a content calendar:
“I call it degenerative engineering. How can I take AI and use it to distill down what I’m really about? Like capture all of this information that’s already happening around my business and then turn that into concentrated orange juice.”
— Mike Montague
He is describing an inversion. Most teams point the tool outward and ask it to manufacture more. You point it inward and ask it to compress what you already have.
Patrick Stox says almost nobody bothers to ask their own experts
Patrick spent six and a half years at Ahrefs before going independent. When I asked him what the frontier looks like now that publishing at scale is table stakes, his answer was not a tool:
“Message five people in your company that are actual experts that know these things are in your industry and like have them, you know, answer some questions, get some insights, get some stories, experience, real experience. And I don’t see people really doing that.”
— Patrick Stox, Unscripted SEO
Five internal interviews a week is not a heroic content operation. It is a recurring calendar invite. The reason it works is the same reason it is rare: it requires talking to people, and the tooling has made it feel optional.
Being retrieved is not the same as being cited
This is the part of Patrick’s answer I keep coming back to, because it reframes what “ranking” even means now:
“we see the same thing in the AI search system so like you get a bunch of pages might be like 90 pages retrieved and then like 13 of them get cited or 30 of them get cited. Well, why were some retrieved and but not cited? And like that’s a that’s a really good question to look into. And it’s usually like you were relevant enough, but like you didn’t quite say the right thing or in the right way.”
— Patrick Stox
You can be in the retrieval set and still lose. The gap between retrieved and cited is filled by specificity — a figure, a named person, a dated example. Those are exactly the things a conversation produces and a prompt does not.
How we actually run it
- Record the conversation before you plan the content. We have not started a keyword cluster from a tool in a while. The seed terms come out of what a guest actually said, because that is where the long tail lives.
- Keep the raw transcript. Cleaned transcripts get condensed, sometimes heavily. If you are going to quote someone under their own name and face, quote the unedited record.
- Strip the host before you attribute. On a long interview the host often supplies more of the tactical content than the guest does. Counting keyword matches across a whole file will quietly credit the interviewer.
- Ship the guest their own version. The interview is worth more to them than it is to you, and that is what makes the next one easy to book.
The part that does not scale, on purpose
Mike’s framing of the two available strategies is the cleanest I have heard. One is to try to be the single company an assistant names, which means getting your material everywhere. The other is smaller:
“winning and deserving the attention of a a small targeted group of people, which is just really good SEO and and marketing. It always has been, but it’s even more important. Like the the hacks are are done because there’s no traffic to be hacked.”
— Mike Montague
There is no traffic to be hacked. So go get something to say.
Quotes are verbatim from the raw Riverside transcripts, trimmed only for false starts and filler. Full episodes at Unscripted SEO.