🎧 Listen to this episode
Mike Montague opened this episode by refusing to be Batman. Jeremy Rivera had just read that telling yourself you are Batman makes you do the task better, and offered it up. Mike countered with Iron Man — and it turned out not to be a bit. It is his entire operating thesis for Avenue9, the firm he runs helping small businesses use AI in marketing, and for the Human First AI Marketing Podcast he hosts.
Iron Man, not Terminator
The Terminator approach, as Mike describes it, is building AI agent robots to go out, kill your leads and bring them back to you. The Iron Man approach is wrapping yourself in technology — Jarvis in your ear surfacing an insight you did not already have — and going out as a superhero version of yourself. He is entirely on the second side of that line.
The pain point he solves sits underneath it. AI is now so big that no one person can track it, least of all a small business owner who is not in tech or marketing. Marketing is also too big for one person to master — you cannot be an expert in SEO and social and content and inbound and direct mail simultaneously. Mike bridges both gaps for people who heard they could use AI for their business, opened Claude, and immediately got asked to set up an integration they do not know how to build.
The C+ problem
Here is the argument that makes the rest of the episode work. Mike thinks AI currently earns a solid C in everything — and that this messes with people’s heads, because if you ask it about a subject you would fail a class in, it looks like a genius. It knows more accounting than you. It knows more SEO than you. But it does not know more SEO than Jeremy does. It is getting a C-minus where a practitioner gets an A-plus.
“A C plus is a really good grade. The problem is that other 20% is the hard part. That’s what we needed humans always to do.” — Mike Montague
That last 20% is emotional awareness, context, the storytelling of your business, the heart and nuance, knowing when to break the rules. As Mike puts it, we never needed AI to cover the easy parts — we had interns for that.
Jeremy coined a name for the failure mode on the client side: Claude Hopping. A client emails you twenty AI-generated documents, forty pages each, announcing they have solved it, and the actual subject-matter expert has to wade through a huge volume of extremely obvious conclusions. Mike had just received a hundred-page marketing plan from a client. The client fully admitted where it came from. The problem was not the tool — it was the assumption that Claude is smarter than both of them.
What AI actually fixed: boring niches
Mike wrote SEO blog content for a garage door company for three years. There are, he says, roughly three interesting garage door stories — safety, security, and curb appeal. He used them up in month one and then had thirty-five more months to fill.
What changed is that he can now get to a C-plus in any subject instantly. On an IT client he can write about cybersecurity processes he does not personally know, speak the language fast enough to be useful, and then hand it to the client to review for accuracy and expert-level detail. That is the actual workflow at Avenue9: interview the founder, the top salesperson, the best client, and turn that into marketing material that has real context and expertise baked in.
Traffic is down and there are only two ways out
The pair traded numbers on the collapse — Jeremy cited a Cyrus Shepard study tracking large blogs where roughly 80% lost about 80% of their traffic, with about 20% of the cohort gaining. Mike’s read is that the market is splitting into two distinct camps, and you have to pick one.
Path one: if AI is only going to recommend one company, be that company. Get your material everywhere, a thousand mentions in a thousand places, not just on your own site — SEO on steroids.
Path two, the one Mike prefers: the old-school stuff. Branding, storytelling, an interesting personality. Start small, start local, and earn the attention of a targeted group of people rather than playing the lottery. The hacks are done, he says, because there is no traffic left to hack.
Make your online reputation mirror your real work
Jeremy’s version, borrowed from Patrick Stox: you cannot put enough SEOs on the payroll to change the internet’s opinion of you, so the most effective way to improve it is to fix what people are complaining about. For most small businesses that is embarrassingly basic — put your phone number at the top of the site, put the reviews up, say what you do and do not do, say what it costs.
Mike raised the ceiling on that. Go get better reviews. Make sure every happy customer becomes a documented blog post, review or case study. Capture real human quotes from every good conversation. Document every job.
The old habit of picking one representative case study is dead: that is a drop in the ocean now. If you are doing a hundred jobs a day, there should be a hundred case studies. And the point is not to manufacture an outsized rank — it is to get the reputation and the traffic you deserve. Nobody is going to take your word for “best plumber in Nashville.” Google and the models have far more people and money pointed at figuring out who actually is.
Degenerative engineering
This is the frame worth stealing from the episode.
“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
Most people, in his view, are working the wrong end of the problem — competing on volume of slop instead of asking AI to compress what they already have into a story strong enough to teach other people to repeat. The related move both of them keep circling is original research: two hosts with a hundred-plus interviews each have quantitative material that can be rolled into reports worth citing. Not “here is what one guy in Massachusetts said,” but “here is what a hundred SEOs said.” Jeremy’s summary: quantity has a quality all of its own. It is also the reason he spun up human-certified content — an LLM will never volunteer the anecdote about the time someone wrecked a Google Search Console deployment.
The old-school channels are quietly open
Jeremy’s direct-mail test at Raven Tools: he looked up agency addresses and hand-wrote letters asking for an opinion to include in an article, offering a graphic to post in return. Ten letters, ten links.
Mike extended it into something anyone with a book can run. Everyone can put a book on Kindle this afternoon, which is exactly why doing so accomplishes nothing. But print ten copies through Amazon, hand-write a sticky note saying “Bill, I think chapter three would be extremely relevant to your business,” and mail it to a prospect. He puts that at 80%-plus engagement, because nobody receives a printed book, nobody receives a handwritten note, and almost nobody demonstrates that they understand the recipient’s business well enough to name the right chapter.
Jeremy’s adjacent play is cross-pollination between the real world and the site: run a seminar at your entrepreneur center, hold a trash cleanup, tow a branded trailer to the park and screen free movies like the roofer he talked to. Then — and this is the part that gets skipped — make sure it comes through the looking glass and shows up online. Rob Bonham described a client doing genuine charitable work that never once surfaced on the web. Documenting it is not self-aggrandizement; it encourages other businesses to do the same.
Entity optimization was never a new thing
On why SEO did not die: Jeremy’s framing is that LLMs are the broader channel and SEO sits in the middle of the Venn diagram, not off to one side. When a model does not know something it runs a tier-one search, then a tier-two, then a tier-three. That is SEO, three times, plus the knowledge graph. Entity optimization — the work Kalicube, Cyrus Shepard, Garrett Sussman and Grant Simmons have been doing for years — is core SEO. We are optimizing the same entities inside a different ecosystem.
Mike’s branding-side translation is the same idea: if you say a thousand random things nobody remembers any of them, and if you say one message, one brand, one promise a thousand times, people remember that.
How he talks clients out of publishing at volume
Jeremy passed along Rob Bonham’s question — when a client insists on publishing AI content at scale, what is your actual process for moving them toward fewer, genuinely differentiated pages without losing the account? Mike’s answer is short: he does not do it, and there are plenty of other people who will. What keeps the account is the reasoning.
“Why would it cite your version if it can already write one that doesn’t need to be cited?” — Mike Montague
If AI can already write your blog post, it already has that knowledge. So the real question is what you could produce that adds to the model — tells it something it does not know. It was trained on the internet and has very little context about you, what happens in real life, or the weird edge cases where the rules break. A thousand blog posts is not a bad idea, he says. The question is how you make a thousand good ones.
The sci-fi turn
Mike’s pass-it-on question for the next guest, which he has now asked over a hundred people: will we reach some kind of superintelligence, and what does the world look like after it? Jeremy’s answer ran a meta-analysis of the fiction — a very low percentage of stories about genuinely self-cognizant AI end well for humanity — and then noted we are already past the dates of most of those futures, and the implications turned out dangerously mundane.
Mike thinks Minority Report is closer than Star Trek, minus the precognition, and points at the Jetsons problem: we do have flying cars, they are called helicopters, and the physics make them loud, dangerous and not worth it. Some problems AI simply cannot solve, so the future will be neither utopia nor dystopia but a messy mix.
Jeremy’s closer went to Dune, and it is the line worth keeping:
“It wasn’t the machines that conquered man, it was man that used machines to conquer man.” — Jeremy Rivera
The Butlerian Jihad was not a war against intelligent robots. It was a revolt against three rulers who uploaded themselves, became machine gods, and subjugated everyone else while the automated society kept everybody comfortable.
What Mike actually does
Avenue9 works in three lanes. Strategy — clients who love AI but know nothing about marketing, and clients who love marketing but know nothing about AI. Systems — the tooling and process, since staying on top of what launches every day is a full-time job by itself. And his favorite, scalable stories — interviewing the people who hold the context, then feeding that into the AI systems that execute. He is currently walking a financial planner through a book chapter by chapter each week, with a custom GPT trained on his voice that still needs the actual stories to work from. SEO and AEO are part of that, but they are components of a marketing plan, not the offering.
Connect with Mike
Find Mike at avenue9.com, on LinkedIn (his preferred channel), and on the Human First AI Marketing Podcast on YouTube.
Listen to the full episode on Castos or watch on YouTube. For more conversations like this, explore the Unscripted SEO topical hubs, try the keyword-research tool at SEO Arcade, or see how these interviews become podcast-powered link building.
