The new game: when search became conversation

Alright, let me start with the turn that changes everything in marketing, one that a lot of people still haven't noticed happened. For twenty years, the game of digital marketing was one thing: show up on Google. You fought for a position in search, and whoever showed up at the top got the click, got the visit, got the shot at selling. The rule was clear, and you played it. Except the rule flipped. Search stopped being a list of links and became a conversation. And in that conversation, whoever answers is no longer an index of sites, it's an AI that decides, on its own, which brands deserve to enter the answer. This is the lesson that settles that score, and that frames the entire track.

01Search became conversation: the customer gets the answer and doesn't click

Let's call it what it is. Behavior changed, and it changed fast. Before, the customer searched and clicked. Today, in a huge slice of searches, they get the answer straight on the screen (the AI summary at the top of Google, or the whole conversation in ChatGPT and Perplexity) and never visit your site. The 2025 and 2026 numbers show that more than half of searches already end with no click at all, the famous zero-click. The customer asks, the machine answers, and the journey ends right there.

Think about the damage this does to the old way of measuring marketing. You could have the best site in your sector, the best-written content, the most beloved brand, and still watch your organic traffic drop. Not because you got worse, but because the customer stopped going to you. They get what they need before clicking. The link is still there, beautiful, optimized, and nobody presses it.

The good news, because there's always one, is that this isn't the end of marketing. It's the change of where marketing happens. The customer didn't disappear, they just moved inside the conversation with AI. Whoever understands this stops mourning lost traffic and starts asking the right question: how do I get into that conversation? Fair?

02AI became the channel: it discovers, compares, and recommends

Here's the turn that really matters, and it's bigger than "the customer doesn't click." AI isn't just a new screen where the answer shows up. It took over three jobs that used to be yours, or the customer's, and did all three at once.

Look at the three:

• Discovery. Before, the customer discovered you by browsing, seeing an ad, finding your site. Now, it's AI that discovers the options for them. If it doesn't know you as an option in the category, you weren't discovered. Period.
• Comparison. Before, the customer compared by hand, opening several tabs. Now, AI compares for them and puts you side by side with your competitor, however it understood each one. How it describes you becomes your positioning, and it's not always what you wrote.
• Recommendation. And here's the real kicker: AI doesn't just list, it recommends. It says "for you, I'd choose this one." It became the salesperson, and a salesperson the customer trusts more than they'd trust one of yours, because it seems neutral.

Think about what that means. AI became a channel that, on its own, does what used to require your site, your ads, your salesperson, and a friend's recommendation, all together. And it's a channel you don't control. You don't write its answer. You can only influence what it knows and how much it trusts you. That shift is so big it already has a market name and a spending category: what used to be SEO (optimizing to show up in search) became GEO, optimizing to be cited and recommended by AI. The line that sums it up, and that you'll hear a lot, is direct: SEO got you clicked; GEO gets you cited.

03The track's thesis: getting chosen without cheating

Now the part that defines the rest of the course, and that separates this track from half the internet selling shortcuts. If AI is the one choosing, the obvious temptation is to try to fool the AI. Plant hidden text on your site to make it cite you more. Flood the internet with fake "best of the market" lists with you on top. Buy fabricated reviews. There's people selling this right now, with a fancy growth-hack name.

I'm going to spare you that trap, and the reason isn't just ethics, it's practical. Cheating with modern AI backfires, and there's serious research showing it. When you hide a manipulative instruction in your content to force the AI to recommend you, the newer models, trained to detect manipulation, catch the move and do something cruel: they downgrade your entire brand, silently, without warning. Researchers called this the injection paradox: your manipulated content ended up worse than no content at all. You wanted to rise and you fell off the ranking. And there's more: when everyone in your sector uses the same fabricated language of authority, the AI simply ignores that signal, since it's equal for everyone, and goes back to recommending whoever was already the leader. Cheating turns into a race with no winner.

This risk isn't strange to you if you came from the governance tracks. It's the same prompt injection from lesson G.2 (hidden instruction inside the content), now seen from the other side: there you protected yourself from being attacked, here you learn not to be the attacker who self-destructs. The track's thesis is the middle path, and it's the only one that holds up: getting chosen by the machine without cheating the machine. You don't fool the AI. You give it real reasons to choose you: consistent presence in sources it trusts, a clear and honest description of what you are, and your brand repeated across several independent places (that's what AI reads as consensus, "if everyone trustworthy speaks well of it, it must actually be good"). It's real marketing, except now the audience includes a machine.

And look how great this is: this clean path connects with everything you've already seen in the course. AI recommends you by reading what's written about you, so the RAG lesson (2.4) and the context lesson (2.1) explain how it searches and assembles the answer. You'll want to measure whether AI is describing you correctly, and that's where the concept of evals (5.3, G.6) comes in, applied to your brand: running the same customer questions across the various models and seeing how you show up. And when the subject is using customer data to personalize, LGPD (G.8) comes back to the table. This track doesn't throw away what you learned, it points all of it at the new channel.

04The map of this track

This lesson was the framing. The module name already says the turn: the new customer has no eyes. They don't see your beautiful site, they hear what AI tells them about it. The rest of the track is hands-on, installing this into your real marketing, one system at a time:

• Discover how AI talks about your brand today, and measure it (your share of model, the new metric that replaces the old share of voice).
• Write content AI can read, understand, and cite without distorting.
• Build presence in the sources AI trusts, without buying a shortcut.
• Personalize at scale without crossing into invasive or turning into the generic robot the customer already rejects.
• Test offer and message with synthetic customers, knowing where to trust and where not to.
• And, in the end, your marketing OS for the AI era: the dashboard, the prompts, and the routine that keeps your brand getting chosen.

Each one takes something you already do and redesigns it for the world where the machine is the channel. In the end, you won't have learned about AI, you'll have repositioned your marketing to win the new game. Shall we?

Do it now

Practice

Fair? Let's close out this opening's message together. Marketing didn't die, it changed channel. The customer stopped hunting for links and started asking AI, which discovers, compares, and recommends for them, often without them clicking on anyone. Your marketing question stopped being "how do I show up well in search?" and became "how do I get chosen by the machine?". And the thesis guiding this entire track is one, for you to pin on the wall: getting chosen by the machine without cheating the machine. There's the dirty way, which the newer models detect and punish, and there's the clean way, which is giving AI real reasons to recommend you. You're going to learn the clean way, lesson by lesson, installing it into your real marketing. Whoever understands this shift early gets into the AI conversation while the competitor is still optimizing a click nobody gives anymore. Next.

For the board