An assistant can write you a competent list in four seconds. It has never lost a client’s rankings on a Friday afternoon and had to explain it on Monday.
This site exists because of a conversation. I was interviewing Mike Montague of Avenue 9 about AI in marketing, and I said out loud on the recording that I had had the idea two days earlier: I have well over a hundred interviews with practitioners, and the thing those interviews contain is the exact thing a language model structurally cannot produce.
“it’s not hard for AI to come up with a bullet list, but AI will never have that anecdote of having messed up somebody’s GSC deployment or their Google My Business page.”
— from the Unscripted SEO interview with Mike Montague
Mike’s response sharpened it into something more useful than a complaint.
The feedback loop is the missing organ
“It’s it doesn’t care and it never really gets the feedback on whether those ideas worked or or not, which is another inherent problem to all of that along with the the echo chamber.”
— Mike Montague, Avenue 9
That is the precise defect. A model can describe a tactic beautifully and has no idea whether it worked, because nothing ever came back and told it. A practitioner who tried it on a real site with real money attached knows within a quarter. Those two outputs read similarly on the page and are worth completely different amounts.
What Mike thinks it actually needs from us
“needs human conversations, documentation and transcripts of real human conversations. It also needs original research.”
— Mike Montague
And the reason two people with a microphone are a research instrument rather than a hobby:
“you and I have both interviewed over a hundred people with… quality interviews and the quantitative approach that we can now then roll up into research reports and other original stuff that can then be cited and be more interesting than just like this is what some random you know dude in Massachusetts said. it’s like no this is what a hundred AI marketing people or a hundred SEOs said about this.”
— Mike Montague
Patrick Stox says the same thing from the other side
Patrick spent six and a half years at Ahrefs. His answer to what separates content that gets cited from content that gets skipped had nothing to do with writing quality:
“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
He is describing a production method, not an editorial standard. Interview the people who know, then write it down. It is what journalists have always done and what most content marketing quietly stopped doing when it got cheaper to summarise other people’s pages.
What human-certified actually has to mean
The phrase is worthless if it just means a person typed it. Here is the standard this site holds itself to, and it is mostly about sourcing rather than authorship:
- Every quote comes from a raw transcript, not a cleaned or summarised one. Cleaned transcripts get condensed, sometimes by more than half, and nothing in the file tells you which kind you are holding.
- Trimming is allowed. Rewriting is not. You can cut a false start. You cannot tighten someone’s sentence and still put their name under it.
- The host gets stripped before anyone is credited. On a long interview the interviewer often supplies more of the tactical content than the guest does, and keyword-matching a whole transcript will hand a guest words they never said.
- A number carries its source and its date. A figure lifted from a summary is exactly as unsafe as a quote lifted from one.
- When a transcript garbles a word, the quote gets cut before it. Not corrected, and not printed. Correcting is a guess about what somebody said.
The unglamorous part
None of that is difficult. It is just slower than generating, and slower is the entire moat. Mike’s framing of the alternative strategy is the one I keep quoting back to people:
“instead of trying to find a shortcut Build the long way. and by the time you get there, you win both of those races”
— Mike Montague
Go have a conversation with somebody who has done the thing. Record it. Keep the recording.
Quotes are verbatim from the raw Riverside transcript of the episode, trimmed only for false starts and filler. Full conversation at Unscripted SEO.