Introduction

host—JORDAN CHENEVIER-TRUCHET:
Hey Kieran, welcome to the show.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Hey, thanks for having me on. Excited to be on.

host—JORDAN CHENEVIER-TRUCHET:
Kieran, I think we have a lot to talk about today, but mostly around AI — because you've been quite loud on AI, the way you use it, your thoughts on it. I really want to dedicate this episode to that. What's interesting in your case is that you have the two levels: the strategic one (AI marketing leader), and the operator side. Especially coming from your background as a software engineer. You were a software engineer, correct?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Yeah, I was a failed software engineer. I did computer science at university, did some years coding — Linux, Unix, Perl, Python. Going back to the command line through Claude Code has actually been great for me. I love working through the command line.

Killing the middle tier of marketing orgs

host—JORDAN CHENEVIER-TRUCHET:
You've said publicly that AI is killing the middle tier in marketing organizations — people who aren't exceptional craftspeople or highly technical. The thing is, most CMOs and VPs of Marketing have built their org around exactly that middle: headcount, job levels, team design. So how should they consider changing the marketing organization?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
I was speaking to a founder recently who uses a lot of agents in his business. His take is probably right for where we are today: if you have someone who's mediocre at their job, agents can take that work and do it just as well. They cannot do the work of a true craftsperson, a real domain expert. Maybe over time they will, but that's where we are today.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
How does it apply to marketing? What happens in all organizations is that we flatten the org structure — people become much more autonomous and can do much more work because of AI. Marketing is unique among go-to-market teams. In sales, sellers all have pretty much the same role. SDRs, BDRs, AEs — same with customer success, support, even engineering. Marketing has a whole slew of niche teams: product marketers, brand marketers, demand marketers — all with their own domain expertise but not really cross-functional.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
What I think happens because of AI: you'll have more generalists who can do all of marketing, with agents and AI as part of their team. The org gets flatter. Probably less middle managers. People managing many more individual contributors. Those individual contributors will be generalists with agents and assistants helping them across all facets of marketing. And that's already happening — spend any time with AI-native companies, that's how they're set up. Not like traditional marketing teams in tech.

The most dangerous employee right now

host—JORDAN CHENEVIER-TRUCHET:
For most marketing leaders, they ask their team to adopt AI to be more productive — but people internally start to wonder, "am I just training AI to replace me?" How do you manage that tension?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
There's always going to be a world where you need to have great taste. The most dangerous employee within any organization right now is someone with a small amount of domain expertise and a high amount of AI agency. They're not very good at their job, but they're really passionate and shipping a lot of AI work. They're the most dangerous people in any organization, because they're creating a real amount of slop.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
We're really in the messy middle right now. AI has given people superpowers — they can ship code, ship memos, ship an infinite amount of stuff. But without real domain expertise and understanding why you're using AI, what outcome you're trying to drive, it just creates more and more noise. Everything looks like an AI problem: "I'll use AI for this, ship more of that." But are you really driving more outcomes? Are you really making the business more successful?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
The best people in a company are the people who have real taste, real domain expertise, and high AI agency. They're the ones creating most of the momentum today. As for the question "am I training AI to replace me" — AI is more on personal accountability than the company telling you to do something. It's your choice to make. AI is going to happen whether any of us like it or not. We have very little control over it. So all you can do is try to make the best choices to be the best version of yourself in a world where AI is prevalent.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
What I'd be thinking about: in a world where AI and agents augment what I do, what is the real skill, the real craft, the real domain expertise I can bring to make myself valuable? And one more thing: model capabilities are moving really rapidly, but integration into how companies work is moving really slowly. There's a lot of opportunity to be the one who figures out how to integrate AI into your team, into your business. That's a skill set most companies need.

Why AI productivity is harder to measure in marketing

host—JORDAN CHENEVIER-TRUCHET:
You've said AI projects live or die on data quality. You shared that at HubSpot you had an AI personalization workflow for booking meetings, and it took 3-6 months to get the prompt right. Most B2B companies have fragmented data, messy CRM. What's the minimum viable data foundation?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Every time we've improved the quality of our data feeding the AI to personalize emails, chat, anything — the increases in conversion rate have been pretty startling. Every time we integrate a new dataset and unlock more personalization, we see a 3 to 500% increase in conversion rates on email, which has been around forever. The minimal viable version is whatever you need to elevate your AI experience above the average in your industry. Data sources depend on the vertical and your customer.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
You want a data repository with clean, enriched data the AI can easily use. The example you mentioned was AI prospecting — we've done that all the way through the funnel. AI has taken over a lot of our prospecting and we've seen huge results. Each time we've given the AI a better dataset — internal CRM data, external from various sources — we continue to see real performance increases.

host—JORDAN CHENEVIER-TRUCHET:
You mentioned earlier that measuring AI in software engineering is easy — you have lines of code, you can correlate that to productivity. In marketing, it's harder. What do you follow to see you're on the right track with AI?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
It's a good question. I don't think we've come up with a de facto way to measure all the different teams. We do show-and-tells, we have a singular Slack channel where people post what they're doing with AI integrated across go-to-market. There are three different ways AI usage shows up: my personal usage, my team usage, and how I integrate it into the customer experience. We measure a lot on the customer experience — AI prospecting, sales agents, support agents, success agents — based on business performance.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
The harder one is visual usage. If a product marketer is using AI, what should we expect in terms of output? Five positioning statements per month vs twenty? It doesn't make sense. This is why I think if you're going to lead teams through this next iteration — and you're best in marketing — you have to be an AI-native leader. There's so much nuance in understanding why and how your team is using AI, whether it's driving the right outcomes. There's no clean metric. You yourself have to really understand AI, dig into the details, and provide good feedback. Marketing more than probably any role should have leaders who are quite AI-native, probably in the weeds doing a lot of stuff themselves.

How to hire AI-native marketers

host—JORDAN CHENEVIER-TRUCHET:
How do you make sure when you recruit people that they're already using AI the right way?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
You can give people tasks, and you make the task pretty impossible unless they actually understand how to use AI. In the past, you'd hire a product marketer with a positioning statement exercise. Now you can say "rebuild the website" as part of the task. That's a really easy thing to do with AI. And if they rebuild the website, you see how they bring positioning to life across pages.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
You can give tasks that are a forcing function — if you don't know how to use AI for this task, you'll think it's an incredibly impossible thing to do. Within the interview process, you can also talk to someone and understand whether they use AI. It's not something you can fabricate if the person on the other end actually knows how to use AI. So: the test you give, the interview process, and the portfolio of work — show me real examples of problems you've solved, times where you had high agency and used AI to solve problems traditionally handed off to others.

Creator-led marketing & AI mascots

host—JORDAN CHENEVIER-TRUCHET:
You've said creator-led channels are becoming the dominant B2B brand asset. But most B2B companies don't have a Kieran Flanagan at the top — their executives are cautious about going public. What should they do?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Setup: in 2021 we bought The Hustle. We thought B2B would look like B2C over time — usually B2B is 3-5 years behind B2C. Back then, B2C attention was shifting into creators and individuals. Branded channels like paid CAC were going up. Blogs were getting disrupted by AI. The channels growing were personality-led — YouTube, newsletter, TikTok, Instagram — where the person is at the forefront. We weren't really thinking about AI back in 2021. We bought The Hustle and built a media network — podcast, YouTube, all assets led by people versus brand, but part of the HubSpot network.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
That's been accelerated with AI. People want to connect with people because they trust people, and a lot of branded content gets lost in the AI ether. There's just so much content out there. We're seeing that happen in B2B today. Your question — can every company do this if they don't have a good creator within the org? It depends on the vertical and the customer.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Here's what I think will happen: most brands will have an "AI mascot." That AI mascot will be the face of their brand. An avatar — a synthetic — and people may hate this, but it will be a synthetic avatar that looks very lifelike. It will bring your brand to life across these personality channels. Your content team is the brains behind that mascot. You might think that's a wild idea — it's already happening in B2C. A lot of the biggest Instagram accounts and the fastest-growing ones are AI avatars. You can't tell the difference, but they obviously disclose it. Same on X. You always want to disclose it.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Will people consume content from AI? Already proven — they are, whether they know it or not. There was a study released — I think NYT or another large media org — a blind study where they gave people AI content and human content. The AI content won. Humans objectively, when they don't know it's AI, prefer the AI content because it's better than the average human content. Will there be a moment of backlash? Probably. But it's going to happen. Pull ahead three years and I suspect that some of the biggest creators in B2C and B2B will be avatars with teams of humans behind them.

Kieran's daily AI workflow

host—JORDAN CHENEVIER-TRUCHET:
Before you started recording, you said showing your day-to-day AI use cases is like asking you to show your day-to-day computer use — hard to even know what to highlight. Can you elaborate?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Early on when I talked about AI on podcasts, people asked: "show me what you do in Claude." Back then, lots of ChatGPT usage. I've migrated most of my usage into Claude. I'd say things like "here's a memo, red team it, give me all the reasons it won't work" — reverse engineering, like Amazon does. Then you fix problems before doing the thing. A lot of strategy work. I had an agent doing first-principles thinking, then frameworks. But over time, the models got so good I don't think about how I'm using it anymore. It's just part of how I work.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
One tip I can give: I have a Claude project for everything I'm working on. Every document, artifact, meeting note goes into that project. I think of them as mini brains. For a project, I have a mini AI brain in a Claude project. I go to Glean (we use it internally), pull all my notes, meeting notes, Slack messages — and feed them to Claude. We've connected to Glean's API so I can say "before you answer, retrieve the latest updates in these docs." Everything I can put in, I put into Claude in that project as a little brain. It helps me work through things in such an expedited way.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Other tips on tools: I built all my decks in Lovable. Recently I did a pitch off internally with a Claude Code-themed presentation styled like a terminal. Lovable lets you create any style. Claude is also good for presentations. NotebookLM is great. Anything using NanoBanana works well. I've passed the part where I have defined workflows. Instead I'm just working on a task and Claude is a partner working on it with me.

host—JORDAN CHENEVIER-TRUCHET:
You generated the deck in Lovable in terminal style — but can you really get these tools to do presentations linked to your brand book?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Yeah. A UX designer did an awesome internal presentation recently for a leadership team. I asked Lovable to copy it for me. One shot. Completely replicated the design. If you give it your brand styles, colors, palette, it can replicate your presentation really well. The reason you might not use Lovable for external conferences is you'd need a PDF. But anywhere you can do something using a website for your presentation, the creativity in Lovable is great.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
One thing I'd push on: should we even be doing things the way we've always done them? All our decks are uniformed, same brand styles and templates. Maybe that's not the way things should be done. AI lets you be creative in ways you couldn't before. What I see a lot of people doing is "how do I get AI to replicate the things we did before?" That's the worst use I see — "enter your domain and we'll replicate your writing style." But your writing style is probably average. Why would you want to replicate it? AI can do a hundred times better than what you do. Why force it to do something that's bad and generic? I don't think people use AI in the right way. They try to do more of what they already do, not elevate what they do.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
We're in the era of show, don't tell. Anytime you can prototype something and show it, it's way better than a presentation. Way better than a memo. Not always doable, but trying to think about how AI can let us do something different — not just replicate what we've done before with AI faster.

Live demo: Kieran's Claude Code content system

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Let me show one of the systems I built in Claude Code. The content team works automatically — I give it a content strategy and it creates a certain amount of content each week. Lessons that apply to any model, any chat: this works just as well in Claude desktop.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
The team is split: an onboarding process creates an audience profile, then a writing style. The writing style can be based on anyone — give the team a name and it creates a style. If you're not a great writer, don't tell it to recreate your style. Tell it to elevate your style.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
People say "this is copying others' writing styles." There's a great copyright framework called copywork from the early 1900s — that's how people got really good at copywriting. Before computers, they'd literally take great writers' work and write it out by hand, letter by letter, to get used to the beats of how someone creates great content. Then you enhance with your own style. That's the writing-style step in my system: get a style card as a foundation, then build your own voice on top.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Then research skills. They create a queue, stack-ranked. An agent provides feedback to the queue and constantly improves it. I think of content in modular blocks: a hook, an intro, an enrichment skill, individual skills per platform. The big thing if you're building real systems: it has to have feedback loops. A performance agent scrapes the web, looks at the content created through the system, integrates analytics, and another agent updates the entire system based on learnings.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
An orchestration skill ties everything together. Rather than remembering all the skills, I just say "let's create content" and the orchestration skill knows all the others. It feels like a little app.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
What's it doing? It's connected to X, Perplexity, OpenAI, and Firecrawl (a tool that crawls websites). Audience profiles are based on what works for me on each platform — the system looks at all my Substack posts plus analytics and says "based on that, this is a winning profile." That matters because the research skills then go find ideas that look like the winning formula for each platform.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Lessons for anyone building this kind of thing: (1) Context really matters. The profiles get pre-loaded — the system has context on who it's creating content for. Whether you're doing product marketing, content, anything — context drives quality. (2) Prompts matter — and they're based on your domain expertise. Your ability to tell the system how to complete a task really well is what makes it skillful. (3) Feedback loops. Show the system examples of what you wanted vs what you got. It gets much better when you're clearer about the outcome you want.

host—JORDAN CHENEVIER-TRUCHET:
How long did it take to build this system?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
It's based on an app I launched last year. The prompts behind the system took about 9 months of iteration. Building it inside Claude Code wasn't long — coming from an engineering background, system building feels logical. Claude Code is so quick at building systems. There's a slash-insights command — run it and Claude shows recommendations for improving the project. You can just say "implement these" and it does.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Over time you can ask Claude Code: "based on the system we've built and your improved capabilities, what should I change, what don't I need anymore?" It really helps you simplify. The value I add is domain expertise — how to bring this to life in a real content system that adds real value. Claude can build the skills and scripts. I didn't build any scripts. It built all of them.

host—JORDAN CHENEVIER-TRUCHET:
How clear are you on the destination before involving Claude? Or do you involve Claude in defining the destination?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
I start with an idea of the system, then work with Claude. I always say to Claude "are we aligned?" — that's a great thing to do before it starts working. It does the ideation. In this example, I had a ton of source code from the app and gave it the entire source code: "here's what I have. How can we replicate this into a system?" It extrapolated all the prompts to start building skills. Then it's iteration. I added the content queue because the more research it did, the harder it got to figure out what ideas it had created. I added feedback loops because it wasn't always right — needed guidance.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
I'm rebuilding the YouTube part because YouTube is so different from text. It's involved in the entire process from start to finish — sometimes Claude is the dominant partner taking the lead, sometimes I'm dominant telling it exactly what to do. Not too different from how you work with an engineer.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
How do I use the output? I never copy and paste content. I create ideas enriched with modular blocks — a story, a case study, an authority quote, a counter-argument, a visual hook — and then I rewrite from that. The system is very much built to how I work. Now if I were giving it away, I'd add an onboarding agent that adapts the system to the user. AI is very personal, very nuanced to how you want to work.

Where to find Kieran

host—JORDAN CHENEVIER-TRUCHET:
For people who want to follow you and learn from you, where do we send them?

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
I did a LinkedIn post about my AI content system. I literally said as a joke "you can only get it on my Substack." I now have 40 DMs on LinkedIn saying "AI content system." So stop DMing me — you can get it on my Substack: kieranflanagan.io. And our YouTube channel is Marketing Against the Grain — with my friend Kipp Bodnar. You'll see a lot of this stuff there visually, watching us build it.

host—JORDAN CHENEVIER-TRUCHET:
All right, thanks a lot Kieran. Talk to you soon.

guest—KIERAN FLANAGAN—SVP Marketing, AI & GTM:
Thank you.

Introduction

host—JORDAN CHENEVIER-TRUCHET:
Hey Alice, welcome to the show!

guest—ALICE DE COURCY—ex-CMO Cognism:
Hi, thanks for having me.

host—JORDAN CHENEVIER-TRUCHET:
Alice, you're the ex-CMO at Cognism. Can you walk us through your journey at Cognism?

Joining Cognism: From €3M to €100M

guest—ALICE DE COURCY—ex-CMO Cognism:
I was at Cognism for nearly seven years. I joined when we were just under three million in revenue. When I left, fast approaching the hundred million mark. The team went from three of us in marketing when I joined to about 28 by the end. We acquired a couple of brands during that time. The reason I was brought on: create a predictable, scalable, revenue-generating marketing engine.

What Changes From Build to Scale

guest—ALICE DE COURCY—ex-CMO Cognism:
When you're at zero to one or three to four, you have to find out how good you are at problem-solving with little to no resources. You learn things that traditionally wouldn't fall into your role. You build operational infrastructure from the ground up. You'd never do that at a larger organization, but you have to in those early days.

You get good at unscalable things in early stages because that's what gets success. You operate fairly blind on data. You have to be comfortable acting on directional insights. As you scale, the expectation is you're no longer the bottleneck. You need to give that knowledge to your team.

Becoming a First-Time CMO: Re-Interview Yourself Every Year

guest—ALICE DE COURCY—ex-CMO Cognism:
Don't have unrealistic expectations — this doesn't happen overnight. The way I thought about it: every year I was reinterviewing for my job. Whether Head of Marketing or CMO, every year I needed to re-interview for the role. Nobody told me. It was just my mentality, because I had never done that next phase.

Building Trust With the CEO

guest—ALICE DE COURCY—ex-CMO Cognism:
I worked out what mattered to the CEO. I know something matters to James when he's mentioned it multiple times in the same time period of a day. Not within a week — within the same time period. That meant: prioritize getting an answer.

I worked out how he wanted to be communicated with. He was a founder-CEO operating at high intensity. When he asks me a question, give him an answer in the moment. That was where I needed to meet him. Always have the numbers — if he needs an answer, it's because there's a board member or investor.

Marketing the Marketing Internally

guest—ALICE DE COURCY—ex-CMO Cognism:
I wasn't great at this in the early years: be super proactive distilling down what marketing was doing and how it led to growth. Don't be afraid of repetition. Repetition is what builds memory. The more I proactively communicated marketing's role and wins that mattered strategically, the better. He wasn't going to come and ask. It had to come from me.

Building the Team: Don't Hire to Hire

guest—ALICE DE COURCY—ex-CMO Cognism:
I was big on: don't hire to hire. I needed something to break before I'd hire. We were resource-constrained. I wasn't hiring for tenure. I was hiring for attitude and ability to understand the new playbook.

I had a core belief about what kind of marketing organization I wanted. People are a good or bad fit for that quickly. We trained up generalists. As we scaled, they became specialist where the data showed we should invest.

The Structure at €100M Scale

guest—ALICE DE COURCY—ex-CMO Cognism:
Paid acquisition function. Demand generation — my business unit owners, owning targets per region or segment. Other parts of marketing acted as mini specialist agencies to them. Content and SEO split: SEO for capturing demand, content for the demand creation media engine. Customer marketing and product marketing under the same VP layer. Enablement team — video, website, design — a resource across the org. And an incredible project manager who kept everything cross-departmental on track.

Where AI Would Have Changed Things

guest—ALICE DE COURCY—ex-CMO Cognism:
Two categories of mistakes. First: when to in-house, when to outsource. We had a growth-at-all-costs era — just threw people at problems. The problem: you can only keep quality of hires to a certain level when your team is a certain size. At too large a scale, you make sacrifices.

Two examples: paid and SEO. I wanted all the experience and domain expertise in-house. The cost of managing that number of people — ensuring career pathways, work variety — it can't happen at scale efficiently. There's a ton of execution work that's really boring. AI is taking a lot of that away.

Specialist in-house plus an agency or AI for execution. That's where I ended up. Had AI been where it is today, that would have happened sooner.

The Ten-Years-At-Salesforce Trap

guest—ALICE DE COURCY—ex-CMO Cognism:
Second mistake: senior hires. My philosophy was to grow people. That got challenged as the C-suite tenure changed. I tried hiring much higher experience levels. The promised land: it'll be better, you'll manage less, you'll learn from someone better at that domain.

A good example: product marketing. We tried a couple — with no success. Never lived up to expectations or cost.

Just because someone has tenure on their CV does not mean they'll best execute that role. Done at scale, it's detrimental to culture. People see they're hired over with the promise of learning from someone more experienced — if that value isn't delivered, it's an additional layer between you and the work. Career blocker.

host—JORDAN CHENEVIER-TRUCHET:
When does it make sense to hire someone with 10+ years from a big company?

guest—ALICE DE COURCY—ex-CMO Cognism:
I wouldn't. The only exception: someone who went to a build-phase company, stayed two to three years, succeeded, then went back into building. Hiring someone from that enterprise type and putting them into a build phase — if they haven't been in build phase for 10 years — is never going to be successful. Differentiate the manager versus the builder.

Why Most First-Time CMOs Get It Wrong

guest—ALICE DE COURCY—ex-CMO Cognism:
Too hands-off for too long. CEOs say to me weekly: "I just want to get some stuff done. Get into execution mode." That's a critical piece marketers shy away from — the execution work. People are scared to hire marketing leaders because all they think they'll get is a positioning statement, not something that moves the engine.

Set expectations upfront with the CEO. It's not about buying hours — you should be buying an output. I should be delivering a product that helps your business move forward.

The Math That Killed Gated Content

guest—ALICE DE COURCY—ex-CMO Cognism:
I was a full-on advocate of ebooks and gated content. We were the best at it. Then we hit another phase of growth. I split the funnel: declared intent demo requests on one side, content leads on the other.

I needed 25 demo requests to close 1 deal versus 100 content leads to close 1 deal. ACV decay on content leads — half the ACV. Look at next year's targets, think about budget and resources — not just on marketing, on sales to follow up — the math doesn't math. Unscalable.

I knew the first question was: "can't we just do more demo requests?" Yes, but how scalably? That came from building an engaging, authoritative content engine — always on, all the time, to everyone in our ICP, delivering differentiated value. I couldn't do that AND gate content.

Selling Demand Creation to a CEO

guest—ALICE DE COURCY—ex-CMO Cognism:
I asked the C-suite, "how do you buy?" When you see an ebook, do you give your real personal details? You know it's coming — a PDF in your inbox. Does it actually get read? You know you'll get cold-called from an SDR. The answer was always no, no, no.

It's hard to go against the logic that this isn't the reality. The math, the reality, plus surfacing that sales are aligned — those three things generated buy-in.

host—JORDAN CHENEVIER-TRUCHET:
How do you make the math math?

guest—ALICE DE COURCY—ex-CMO Cognism:
Not cold turkey. Turn down the most inefficient content campaigns. Repurpose that money into ungated. Hypothesis showing how I'll scale demo requests in line with the spend that fills the gap.

The terrible conversion rates on content leads worked in my favor — I didn't need many more demo requests to fill the gap. Within two months I had correlation graphs of spend and demo request increase. As you scale, you invest in software — Fibula on LinkedIn ads, Dream Data across the journey. Eventually proper marketing mix modeling at €100M+.

The Four Content Buckets

guest—ALICE DE COURCY—ex-CMO Cognism:
You need a strategy. Topics you want to be known for. Memorability is built through repeatability. Random acts of content, even ungated, won't have impact. Four content buckets:

• Thought leadership — the new story of the day for our industry, broken in a helpful, actionable way.
• Dotted line back to the offering — we still need people to know what we do.
• Product content — not decks. How does the product solve a problem? Voice of customer or third party.
• Social proof — facts and opinions of others, not just us.

And distribution. Understand where your audience hangs out and what formats work for those channels.

You cannot cheat quality. We're never going to be as big as competitors on spend. But we can create truly authentic, free quality content that an LLM can't write and competitors can't do at their scale.

Acquisitions and Brand Integration

guest—ALICE DE COURCY—ex-CMO Cognism:
Strategic reason for the acquisition drives everything. Take Caspr — incredible brand presence in France and we were going into France. We weren't going to kill the brand. PLG motion versus our sales-led motion — we didn't want to cannibalize our SMB audience. Different brand for that reason.

On the back end, fully integrated. Same office, same systems, integrated teams. The Caspr marketers became the French leads.

Two Books and Closing

host—JORDAN CHENEVIER-TRUCHET:
You wrote two books. Give me the sales pitch.

guest—ALICE DE COURCY—ex-CMO Cognism:
The honest ramblings of someone trying to master the CMO role for the first time. A story of two halves — early growth versus later stage are different. The journey from a content lead generation business into a fully bespoke create-demand motion. Anyone in a leadership role for the first time, or building from the beginnings.

host—JORDAN CHENEVIER-TRUCHET:
Where can people reach out?

guest—ALICE DE COURCY—ex-CMO Cognism:
LinkedIn. I'm very active. Put the bell on the LinkedIn posts.

host—JORDAN CHENEVIER-TRUCHET:
Thank you Alice.

guest—ALICE DE COURCY—ex-CMO Cognism:
Thank you.

Introduction

host—JORDAN CHENEVIER-TRUCHET:
Hey Dave, welcome to the show!

guest—DAVE GERHARDT—Founder Exit Five:
Thanks for having me, Jordan.

host—JORDAN CHENEVIER-TRUCHET:
Now everybody's talking about AI. CMOs and VPs of Marketing are scared. There's huge FOMO. The CEO and the board are pushing "how do we double revenue and divide the team by three with AI and agents?" On your side, you talk a lot about events, social, organic, storytelling. What's wrong with you?

The AI Hype vs Reality Gap

guest—DAVE GERHARDT—Founder Exit Five:
I'm not an engineer. This is the perfect time if you came up to marketing from being an engineer. My friend Tom Wentworth, CMO at Incident IO, posts every day about things he's building with Claude Code. He has engineering background — perfect for this era.

I run a community of B2B marketing professionals — CMOs and marketing leaders. There's a huge gap between what you see on YouTube and what's actually happening inside companies. If you go to YouTube and type "Claude marketing," you'll see 20 videos: "I replaced my marketing team with Claude Code in 20 minutes." Then the next one says 16 minutes. There's a huge cash grab and attention grab.

I'm at the intersection of "I believe in AI" but I also talk to lots of people every week who'd be surprised how many people aren't even using Claude regularly. Just using Claude is a revelation for them.

And ultimately, you're still selling to people. Until my AI agent does research for me and buys on my behalf, exclusively talking to your AI agent, humans are still in the buying process. There's a lot in social, community, and events that pays towards humans.

Do You Use AI?

host—JORDAN CHENEVIER-TRUCHET:
Do you use AI?

guest—DAVE GERHARDT—Founder Exit Five:
Absolutely, of course. I am AI. Claude was actually trained on me.

What Stays Timeless: Human Connection

guest—DAVE GERHARDT—Founder Exit Five:
A lot of stuff that can be replaced makes sense. Advertising is a perfect example. It's always been about how many landing page variations, copy, creative, headlines, images you can produce. Ad creation, landing page creation, automating ads — the future is: I should just be able to go to LinkedIn, type in plain text "show these ads to these people," and it just happens automatically.

But on the other end, we just came back from doing this event with 100 people in Arizona for B2B CMOs. Highly curated, lots of offline time, real conversations. There's something about that human in-person connection.

We're not all going extinct in the next couple years. There's a fundamental need to still want to connect. Content creators are filming videos with old-school video cameras. iPods are back. Landline phones are back. People are dying for a phone with no screen, doing retreats, "raw-dogging" a flight. We're trying to find ways to reconnect, be human, be present.

I want to be in both camps — believe in AI, use it to run my business, but not lose sight of humans being connected over the campfire, telling stories.

Marketing's New Superpower With AI

guest—DAVE GERHARDT—Founder Exit Five:
I became a CMO because I was good at marketing. You get promoted, eventually rise to the top, and "congratulations, you're good at marketing — you now manage the whole marketing team." Turns out when you do that, you don't actually do a lot of marketing. You do people management, budgets. Most of my day was with HR, the CFO, the CEO, fundraising. I was like, "man, I just like writing copy on a landing page and making an ad."

What's so cool now: I can have an idea — I'm not a designer, I'm not a developer, I only know basic HTML — but I could go this afternoon and build my own website. The speed at which we can go from having an idea to making something is exciting.

From CMO at Drift/Privy to Building Exit Five

host—JORDAN CHENEVIER-TRUCHET:
You had to start from scratch when you left Drift and Privy. How did you handle that?

guest—DAVE GERHARDT—Founder Exit Five:
I never felt like I went from scratch. When I was doing marketing at Drift and Privy, I started writing about marketing on LinkedIn. I had a podcast and a newsletter. Tale as old as time: if you write about the things you're doing online, you attract the audience of like-minded people.

That led to me unintentionally building a large following on LinkedIn of B2B marketers. It wasn't because I wanted to be famous or have a personal brand. I live in the woods in Vermont. I don't drive fancy cars. I'm never going to be the personal brand guy in front of a Lamborghini.

Patreon, Then Substack: Pricing Content From Day 1

guest—DAVE GERHARDT—Founder Exit Five:
Then it was natural progression — "I have a lot of followers, I'm going to launch a paid version." The closest comparison now would be Substack. Back then it was Patreon. $10 a month. People today are like "Patreon? What the hell is a Patreon? Are you making porn?"

Three to six months in I had a thousand members. That was $10,000 a month in recurring revenue from content. That broke my brain because forever in my mind, $100K in a year was the epitome of success.

That was while I was a CMO. COVID happened, I was home every day. I built that while still CMO at Privy. I had product-market fit before going full-time. My career has just been stacking things on top of each other.

Why Most B2B Communities Die

host—JORDAN CHENEVIER-TRUCHET:
Most B2B communities are like dead Slack groups. What did you do to make Exit Five work?

guest—DAVE GERHARDT—Founder Exit Five:
A lot of communities go to zero. They let them get over-riddled with spam and promotional stuff. One way we solve for that early on is by charging money. People thought I was just an awful arrogant evil capitalist. But money is an amazing forcing function to get commitment. We just did our event in Arizona. The two pieces of negative feedback were from the two people who had free tickets.

The Rules: Charge Money, Stay Niche, Police Comments

guest—DAVE GERHARDT—Founder Exit Five:
Number one, we charge from the beginning. Number two, having a really focused niche — exclusively for B2B marketers.

Number three: I've written maybe 8,000 posts all-time and 40,000 comments. For the first five years, I'm in there every day — writing, commenting, policing comments. If Jordan writes a lazy post, I'm like "Hey Jordan, update your post — it's not very specific." Or "this is clearly you promoting your company." Founder mode. Things that don't scale.

Communities mostly die because they're an acquisition channel, not the product. If a vendor owns it and the goal is nurturing people into a SaaS subscription, engagement drops. Every time. Exit Five isn't owned by a vendor. The community IS the product.

Storytelling as a Competitive Skill

guest—DAVE GERHARDT—Founder Exit Five:
It's hard. It's almost like a friend who just has a way. The "je ne sais quoi." You can't go to ChatGPT or Claude and study a storytelling structure and just print it out. The people that are really good at marketing have a wide range of interests outside work.

Especially in B2B: "let's go see how Stripe does their website." Where I'm like "shoot, check this out — this lady down the street has an amazing sandwich shop here in Vermont and her website is awesome." The confidence to not just copy what's been done in your industry.

Writing With Claude (and Why It Takes Two Hours)

guest—DAVE GERHARDT—Founder Exit Five:
I have a newsletter — exit5.com/newsletter. I use Claude to write it. It takes 2+ hours.

I'm the subject matter expert. I didn't just give it to AI and one-shot it. It's: "I have this idea. I recorded a podcast with Jordan. Here's the transcript. Here's two other things I've written that relate to this." That's where Claude is amazing — like having a creative agency partner. I'm now piecing each section together.

It's not saving me time. It's making the output better.

Personal Brand vs Company Brand

host—JORDAN CHENEVIER-TRUCHET:
How do you manage the tension between writing for the company versus writing for you?

guest—DAVE GERHARDT—Founder Exit Five:
It's easy for me — I run the business. But say you're VP of Marketing at a cybersecurity company. You can't write the same way. Get direct with the subject matter expert. People buy from people. The opportunity is sprinkling in personality and being you. Every business is in the business of social media now.

Metrics That Matter: Revenue, Churn, Engagement

guest—DAVE GERHARDT—Founder Exit Five:
I do look at data. Revenue. Money. Goal: make more than you spend. Indicators: growth in membership, churn. Churn is a killer. Email list growth, engagement, opens, clicks. NPS — are people happy with the thing they're paying for?

In B2B inside a company, MQLs are bullshit. But I run a media company — those ARE our metrics.

F*ck Attribution

host—JORDAN CHENEVIER-TRUCHET:
Do you work on attribution?

guest—DAVE GERHARDT—Founder Exit Five:
Fuck no. We ask every person joining: "how did you hear about us?" They tell us. We export them a couple times a year. We don't learn that much. It reaffirms what we're doing is working.

So much of this comes from over-metricing marketing. At some point you've seen the whole game and you just know — if you do the right things, people will show up.

host—JORDAN CHENEVIER-TRUCHET:
So you don't think about the ROI of content.

guest—DAVE GERHARDT—Founder Exit Five:
Hell no. If we put out good content people like, they will join our newsletter, they will be on our podcasts, they make us look good. If people think we're legit, they will. They're not morons.

Closing

host—JORDAN CHENEVIER-TRUCHET:
Where should we send people?

guest—DAVE GERHARDT—Founder Exit Five:
exit5.com is good. Also I'm on LinkedIn — Dave Gerhardt.

host—JORDAN CHENEVIER-TRUCHET:
Thanks Dave.

guest—DAVE GERHARDT—Founder Exit Five:
Thanks for having me.

Introduction

host—JORDAN CHENEVIER-TRUCHET:
Hey, Eli!

guest—ELI SCHWARTZ—Author Product-Led SEO:
Bonjour Jordan, good to be here.

host—JORDAN CHENEVIER-TRUCHET:
Thank you for being on the pod. For people who don't know you, you're famous for having written a bestseller in the SEO space called Product-Led SEO. What is Product-Led SEO?

What is Product-Led SEO?

guest—ELI SCHWARTZ—Author Product-Led SEO:
The story is: I was working at SurveyMonkey — my last full-time job — and I started consulting on the side. A large brand reached out to potentially hire me. I gave them an offer and they asked: "So what are you going to do for us? Write content? Do keyword research on SEMrush or Ahrefs? Recommend links to build?" I said yes to all three.

Then he asked: "Why wouldn't I hire someone in the Philippines, Pakistan, Bangladesh, or anyone on Fiverr to do that?" I knew the answer was: I'm better. You reached out. You wanted to hire me. If you thought you could do it for $50 on Fiverr, you would have.

I didn't have the concrete answer at the time, and I didn't sign that client. But it forced me to think back on how I did things at SurveyMonkey. The way I did things was not just creating keyword-driven content and building links to it. I thought about what the user was searching for and built that asset. I needed a name for it — Product-Led SEO became the name.

What it really is: when someone is searching, they're not searching for something that matches a keyword. If someone searches "best mobile phone for me," they're not looking for a piece of content that says "best mobile phone for me" 60 times. They want the answer. They want a product. Product-Led SEO is building the thing — a piece of content, a video, an experience — that satisfies the answer, not just matches a keyword.

The Product Approach: TripAdvisor and Amazon

guest—ELI SCHWARTZ—Author Product-Led SEO:
TripAdvisor is a great example. When someone searches "review of the Ritz Carlton Paris," they're not looking for content that has the phrase "review of Ritz Carlton Paris" 60 times. They want the review. TripAdvisor's approach back then wasn't to write blog post after blog post. They built a website that satisfies any combination — Paris, New York, Ritz Carlton, Hilton. That's the product.

Same with Amazon. They didn't approach e-commerce the way late-90s competitors did — "let's make a piece of content about what a cell phone is." Amazon put all the effort into building the architecture and structure of great SEO. There are now millions of things sold on Amazon that were never dreamed about when Amazon was created — but they don't need to build a new Amazon for each. They extend the architecture. Even better, they took amazon.com and extended it to amazon.fr without needing a new SEO strategy for France.

host—JORDAN CHENEVIER-TRUCHET:
Would you say "product" is an umbrella word that includes content, reviews, calculators, etc.?

guest—ELI SCHWARTZ—Author Product-Led SEO:
It includes the thing the user is looking for. On a media site, content is the product. On an e-commerce site, the e-commerce landing page is the product. On a research page, it's the ideas. Product-Led SEO is really better defined by what it is not: content-led SEO — keyword research, create something that matches exactly what the user searches, call it a day.

Rewriting the Book for the AI Era

host—JORDAN CHENEVIER-TRUCHET:
Between when you wrote the book and today, a big thing happened: LLMs, AI, AI Overviews on Google. What would you rewrite today?

guest—ELI SCHWARTZ—Author Product-Led SEO:
A lot of my book was intended never to be obsolete. I wrote the whole book myself — I didn't hire a ghostwriter. Every time I used "Google," I added "Bing," just in case Bing would overtake Google one day. I didn't talk about a lot of tactics, again because I didn't want it to be obsolete.

Much of the book applies today. The big thing I'd change: I'd talk about AI being an answer itself. AI Overviews, Gemini, ChatGPT — they become that first-layer product. If a user is just looking for an answer, they no longer need a website that has done SEO.

The other part: I'm excited my vision for Product-Led SEO is now coming into fruition. The shady tactics — buying 80 million backlinks, spamming keywords — don't work as much anymore. What's left is essentially Product-Led SEO. The only thing I'd add: help the reader understand that experience also happens on an LLM, and you can't recreate it if it's happening there.

Will LLM Companies Need SEO?

host—JORDAN CHENEVIER-TRUCHET:
Would you advise these big LLM corporations to invest in SEO?

guest—ELI SCHWARTZ—Author Product-Led SEO:
I believe they already are. Go on LinkedIn — OpenAI, DeepSeek, Perplexity, Anthropic all have growth marketing teams and people with SEO titles.

Contrary to what everyone thinks based on social media — "Google's dead, everyone will use LLMs only" — I don't think that's close to the case. Those companies know that to reach 80–90% of the online world, they have to do it through Google.

Given the antitrust suit in the US, Google maintained its monopoly. Google will dominate consumer search until something completely shifts. Not just because they're good — they have distribution power: Android, Workspace, Chrome. People say "kids aren't using Google anymore." In the US, many kids get their first device from school: a Chromebook. Google's Trojan horse is "here's your new computer, it can only use Chrome."

So those LLM companies know the way to reach online users is through Google and typical SEO.

The Search Pie Is Getting Bigger

host—JORDAN CHENEVIER-TRUCHET:
People see this as a shift from search to LLM, but actually the cake is getting bigger. People still use Google — they use Google plus OpenAI plus other LLMs.

guest—ELI SCHWARTZ—Author Product-Led SEO:
Exactly. There are many queries Google wasn't a good solution for. Four years before ChatGPT, everyone said Google was dead because of TikTok. TikTok satisfied an area of search Google couldn't — video restaurant reviews, trip plans.

Now LLMs do things Google couldn't. I was using Gemini to figure out why a door in my house wasn't working. I explained the problem, took a picture of the handle, Gemini gave me an answer. I never could have done that on Google. This is a new area of search. By the end I'll probably need to buy a part — that's when I'll go to Google. The top of funnel didn't even exist before.

Where AI Wins, Where SEO Wins

host—JORDAN CHENEVIER-TRUCHET:
You've said AI owns top-of-funnel and SEO wins mid-funnel and bottom-funnel — even with AI Overviews expanding into mid-funnel. Where do you draw the line?

guest—ELI SCHWARTZ—Author Product-Led SEO:
First: I'm not sure GEO is a real, separate investment. I'd continue to invest in SEO. Many big websites are going to get themselves in trouble because they think AI is the only future, and they'll make basic SEO mistakes — subdomains vs. subdirectories, technical SEO that should be structured at the start.

Top-of-funnel example: someone wonders if they have a health issue, like anxiety. That belongs to AI Overviews. They describe symptoms, the LLM says "this sounds like anxiety." Many blog posts cover that, but Google's AI Overview takes the top of the page and gives the answer.

I worked with a mental health company. Their solution was therapy, for anxiety. There are many ways to address anxiety: meditate, listen to music, take drugs, go to therapy. But this company only offered therapy. So bottom-of-funnel was "therapy" or location-based queries. The super top-of-funnel — "anxiety" — doesn't exist for them anymore. AI Overviews handle that. Now AI tells the user "people with feelings like yours could meditate, listen to music, or try therapy." That brings them to mid-funnel: "therapy for anxiety." That's the only place this company can address.

Citations Are the 2026 Version of Backlinks

host—JORDAN CHENEVIER-TRUCHET:
The data shows the top SEO results are 80% the top LLM results. So you should still invest in SEO even to appear on LLMs. Agree?

guest—ELI SCHWARTZ—Author Product-Led SEO:
Yes. I talk to so many companies that want me to recommend them for GEO/AEO. Their answer fits two buckets:

• Content directly matching what the user prompts. If they directly match, they show up. But we don't even know what the prompts are if they're long-tail.
• Citations. Citations are the 2026 version of backlinks — and I don't think backlinks have mattered that much in the last couple of years. If the whole GEO play is getting citations, is that really new?

Beyond LLMs, AI is trying to think like a human. You can spam citations all you want, but if AI thinks like a human, it also thinks "which of these citations matter?" Most brands that get cited are cited because they're brands. Build a brand users actually like. That's what the engines are looking for.

I asked John Mueller in Zurich whether Google uses any data from Google Maps and Street View. He said of course they do. If you have a restaurant and Street View goes by, Google Maps tells you if a store is busy. If I were an LLM with access to that data, I'd use it to recommend the pizza shop that's always busy at lunch. You can build all the citations you want, but if Google knows no one goes there, it doesn't matter.

The Five Stages of Awareness vs ToFu/MoFu/BoFu

host—JORDAN CHENEVIER-TRUCHET:
Are you familiar with the Five Stages of Awareness framework?

guest—ELI SCHWARTZ—Author Product-Led SEO:
Yes.

host—JORDAN CHENEVIER-TRUCHET:
Instead of ToFu/MoFu/BoFu, with the Five Stages you focus on Problem Aware. Problem Aware is "I type my symptoms." I don't know any solutions — I just know my problem. I go to an LLM because everything is foggy. But when I know exactly what I need — Solution Aware, Product Aware, Most Aware — I just go to Google.

guest—ELI SCHWARTZ—Author Product-Led SEO:
Exactly. I had this experience with one of my first consulting clients, Mixpanel. We did all the SEO stuff but none of it converted. Mixpanel is expensive and very sticky. People landing on the content were very early in the journey. The right way: what do we do at this very high point of the funnel to bring them to the next part? Maybe retarget. Maybe a webinar. Not a big purchase.

How to Vet GEO Agencies

host—JORDAN CHENEVIER-TRUCHET:
What's one question a CMO should ask GEO agencies?

guest—ELI SCHWARTZ—Author Product-Led SEO:
Start with the assumption that they probably don't know what they're doing. Agencies only selling AIO/AEO/GEO probably don't. If they have a track record of successful clients in SEO and now show how they've adapted to AI, that's more trustworthy.

Same with tools: look at the people who made the tool. Do they have a track record? If they just "found a niche, everything has switched, let's make money while people don't know what's going on" — avoid.

The basics of AI visibility are the same as SEO. So quiz them on traditional SEO. It's the Wild West with AI — everyone's making things up. I see data studies that completely conflict and they're both right because the data is too vast.

Convincing the CEO Not to Over-Invest in GEO

host—JORDAN CHENEVIER-TRUCHET:
How do I convince my CEO not to fall into FOMO?

guest—ELI SCHWARTZ—Author Product-Led SEO:
You don't. I find it very difficult to convince people who really believe in something not to do it. If the CEO really believes GEO is the future, fighting creates friction. What I'd do: do it. Focus 10% on GEO and AI visibility, focus 90% on typical SEO. In six months, if you're right, the CEO will say "this seems to be a scam, let's drop it." You can say "I tried to tell you, which is why I didn't put all my eggs in this basket." Now you deserve a bonus.

Budget Allocation: $1M to Spend

host—JORDAN CHENEVIER-TRUCHET:
You have $1M as a CMO. Walk me through the decision tree.

guest—ELI SCHWARTZ—Author Product-Led SEO:
I would not spend it on SEO. First, it's very hard to spend $1M in SEO properly. Second, if you don't already have a brand, do brand marketing first. Most companies should do paid before SEO — it's quick feedback on what works, on product-market fit, on what I call search-market fit. Are people even searching for this? If not, your SEO won't work.

host—JORDAN CHENEVIER-TRUCHET:
Search ads, not social ads?

guest—ELI SCHWARTZ—Author Product-Led SEO:
Search ads, yes. I actually like social ads too. I'd do social ads first, then search ads, then SEO. The biggest bang for buck for any startup with $1M is brand. Hire PR agencies. Get the name out. That trickles down into everything else.

Should SEO Sit in Product or in Marketing?

host—JORDAN CHENEVIER-TRUCHET:
Should SEO be in the product team?

guest—ELI SCHWARTZ—Author Product-Led SEO:
Ideally yes. The levers for getting things done are better moved within the product team. On a marketing team, you write tickets that go to product, then engineering. On a product team, you and your colleagues all report to the same boss and work on the structural assets directly.

Example: making a CMS decision. At larger companies, marketing has little influence — it's product. My most successful SEO engagement was Tinder. I worked directly for a VP of Growth. We had engineers, marketers, designers in our growth pod. Two years of work, and the impact came because I was in the product team.

host—JORDAN CHENEVIER-TRUCHET:
What's the best org structure?

guest—ELI SCHWARTZ—Author Product-Led SEO:
SEO would be a product manager whose responsibility is SEO. Hub and spoke, with the PM at the center. If you're SEO inside marketing, you become a spoke and lose control.

Technical SEO and Link Building Are Overrated

host—JORDAN CHENEVIER-TRUCHET:
You said link building is overrated and technical SEO is over-indexed. What do you mean?

guest—ELI SCHWARTZ—Author Product-Led SEO:
SEO audits come back with technical feedback — 301 redirects, 302s, 404s. If you implement everything, you're unlikely to see much growth. It's like going to the doctor and being told to run 10 miles a week. Maybe you live longer. Will you meaningfully make more revenue? Most honest consultants would have a hard time saying yes.

On link building: there was a time it really mattered — raw PageRank, no filters. For many years now Google filters that out. I had a backlink from WhiteHouse.gov when I joined SurveyMonkey — the Obama White House did a partnership. The link was a 404. I created a page to fix the 404, then linked from it across the site. All those pages started ranking higher. That was 2012.

In 2018 SurveyMonkey got more White House backlinks — they were scraping our research. None did anything. Why? Google got smarter. The White House is authoritative on government, not on customer feedback surveys. So Google said "this is not a relevant link."

That was 2018. They've gotten even smarter. They look at a Forbes link and say "this is paid placement." So just getting a high-DA link isn't something I'd invest in. Same with citations.

The Future of SEO

host—JORDAN CHENEVIER-TRUCHET:
How do you see the future of SEO?

guest—ELI SCHWARTZ—Author Product-Led SEO:
SEO never disappears. I made a prediction that Meta is about to launch its own search engine — their LLM already exists at meta.ai. It will work by pulling Meta data — Facebook, Instagram, WhatsApp Business. They've made big AI acquisitions. Real player.

host—JORDAN CHENEVIER-TRUCHET:
They actually announced an acquisition today.

guest—ELI SCHWARTZ—Author Product-Led SEO:
The world of LLMs will be dominated by companies that were already huge. OpenAI raised a lot, but beyond OpenAI and Anthropic, big tech wins. SEO is always this request for discovery. Search 15 years ago was "go on your computer and type into Google." Next year it might be glasses. That's still search.

The Most Contrarian Take

host—JORDAN CHENEVIER-TRUCHET:
One contrarian take that you believe almost nobody in the industry agrees with you on?

guest—ELI SCHWARTZ—Author Product-Led SEO:
I really don't think anybody should be investing significantly in AI as a separate thing. Invest in search visibility. The way GEO/AEO is framed today is as if it's something separate. It's like 20 years ago when "mobile SEO" became a thing — as if you had an SEO person and a mobile SEO person. Same trap.

AI visibility is just the 2026/2027 version of SEO. Make sure your existing agency knows how to think about SEO of today, which has an AI visibility element. Don't break it off and call it a new thing — that creates a budget category you don't have to spend in.

Closing

host—JORDAN CHENEVIER-TRUCHET:
Fair enough. Thank you, Eli. For people who want to follow you, you've got a newsletter and you post on LinkedIn. Your book Product-Led SEO is still a bestseller — link in the description.

guest—ELI SCHWARTZ—Author Product-Led SEO:
LinkedIn is great. And please read my newsletter. I've been doing it for three years. When I started, no one read it. I just wrote to the world. Now I have almost 14,000 subscribers. Whenever someone asks how I got there, I just kept writing. Honored if anyone reads it.

host—JORDAN CHENEVIER-TRUCHET:
I'll put the link in the description. Thank you, Eli, all the best.

guest—ELI SCHWARTZ—Author Product-Led SEO:
Thanks for having me.