AI Is Reshuffling the Entire Advertising Industry

A while back, I grabbed dinner with some friends who work in ad buying.

Over the meal, one of them asked: Have you noticed that the way we shop has completely changed over the last couple of years?

I asked how so.

He said, before, when you wanted to buy something, you'd open a search engine, click through several pages, and compare options yourself. Now? You just ask an AI — toss it a sentence, and it hands you a tidy answer. You place the order and move on.

Everyone at the table nodded.

That was the moment it hit me: The way consumers look for answers has changed — so the way brands chase them with ads has to change too.

U.S. digital advertising is a $300+ billion-a-year market, accounting for three-quarters of the entire advertising pie. And AI is carving it up from the inside out.

Today I want to talk about where this reshuffle is hitting hardest.

Search: From "Scrolling Blue Links" to "Waiting for an Answer"

Let's start with search, because it's the one being upended the fastest.

Google holds roughly 90% of the global search market. But a survey by Evercore found that 8% of people now skip Google entirely when searching and use ChatGPT instead. Six months ago, that number was just 1%.

That's an 8x jump in six months.

Think about what that means.

In the past, you'd search "best shows to watch this year" and get a wall of blue links to click through one by one. Now you throw that question at ChatGPT, and it hands you a curated list — with reasons, ranked and ready.

Consumers have shifted from "finding options" to "wanting answers."

That's a disaster for websites that live on clicks. And it's an earthquake for everyone who buys search ads.

So what happens to SEO (Search Engine Optimization)?

The old playbook — building backlinks, stuffing keywords, writing for E-E-A-T — was written for Google's crawlers. But now, the thing reading your content might not be a human at all. It's a model. And when that model assembles an answer for the user, it decides whose content gets cited.

That's why companies specializing in GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) have sprung up — dedicated to making sure AI mentions your brand when it answers questions. Put simply, the target of SEO has shifted from "humans" to "models."

And paid search ads (SEM)? ChatGPT, Claude, Perplexity, and Google's Gemini are all starting to carve out ad space inside their answers. Perplexity has already launched "sponsored follow-ups," where brands pay to surface their results first.

Here's an even more interesting shift: going forward, the thing you need to win over might not be the consumer themselves — it might be the AI agent running errands for them.

Google has already shipped a product that can take over your browser and book tickets on your behalf. Payman and Stripe are building payment rails for agents. Browserbase is laying down infrastructure for agents to use.

Your "customer" is starting to become a piece of code. So does your landing page, your visual merchandising — all of that — still work on a piece of code? Whoever figures out the answer to that question first will grab the next wave.

Social and Influencer Marketing: The Biggest Winner Hasn't Truly Hit Its Stride

Google's YouTube, Meta's Facebook and Instagram, ByteDance's TikTok and Douyin — these three account for three of the world's top five ad sellers.

They hold two things: massive user bases and massive data. In their hands, AI is a sharp blade.

But today I want to focus on something underrated: why influencer marketing has never been able to scale.

U.S. brands spent roughly $8 billion on influencers in 2024. Sounds like a lot, right? Compared to the total digital ad market, it's a rounding error. This market has grown at a 26% annualized rate over the past five years, but the base is still small.

Why? Almost every brand I've spoken with says they want to increase their influencer spend. But the money can't get in.

The reason is simple.

When you buy Meta ads, you set a budget, run it for two days, check ROAS, and if it performs, you scale up tenfold tomorrow. Precise targeting, easy scaling, all in one click.

When you do influencer marketing? You find creators, negotiate with agents, haggle over price, grind through contracts, align on creative, monitor the launch, collect data. By the time you've run the whole gauntlet, weeks have passed. And if it performs well? You want to scale tenfold? Sorry — run that entire process all over again.

The appetite for budget is there; the plumbing to absorb it is not.

Companies like Agentio are trying to change that: making influencer marketing work like feed advertising — targetable, measurable, and scalable with one click. If they pull it off, that $8 billion figure will multiply several times over.

So my own judgment is this: the real opportunity isn't in "building another Meta." It's in "turning influencer marketing into a feed ad product."

Retail Media: Amazon Feasts, Everyone Else Pulls Up a Chair

Meta and Google together once captured more than half of all U.S. ad spending.

Now that duopoly is getting bitten from behind by retail media. Amazon's ad revenue is projected to exceed $42 billion this year, accounting for 14% of total U.S. ad spend. Walmart is further behind at around $4 billion, but it's growing too.

Why are advertisers flocking here?

After iOS 14, precisely targeting people on social platforms got a lot harder. But in Amazon's on-site search, when you type "baby pajamas," what shows up at the top is a sponsored slot. Clear intent, direct conversion.

Uber, Instacart, and DoorDash are all selling ads in their apps now. PayPal and Block are building ad platforms on top of their payment data and user pools.

Where there are buyers, there's data; where there's data, ads get sharper; when ads get sharper, more advertisers come. The flywheel is spinning.

Looking ahead, generative AI will let smaller retailers play this game too. Rebuy powers on-site personalized recommendations for DTC brands. Pietra helps merchants find distribution channels with AI. Topsort lets small retailers build their own ad systems. The kind of capability that lets Amazon's recommendation engine drive a third of its GMV is becoming an affordable, off-the-shelf standard.

Connected TV: Linear TV Is Stepping Aside

In 2019, linear TV still accounted for 30% of U.S. TV ad spending. By 2026, that figure is expected to fall below half that.

Where's the money going? Into Connected TV (CTV).

Why? Because linear TV only lets you buy "reach" and "awareness" — you have no idea who's actually watching. CTV can target, measure, and calculate ROI — the same playbook as digital advertising.

AI does three things here: optimize placements in real time, identify high-intent audiences, and mass-produce personalized creative. Tools like HeyGen and Creatify can generate a TV ad with localized voiceover and store-level promotions in seconds.

It all sounds great — but let me temper the enthusiasm. CTV isn't very friendly to small brands yet. The barriers to entry and minimum spend requirements are still sitting right there. This is a battleground for well-funded big brands.

Who Wins This Reshuffle?

I've been thinking about something.

In the Web 1.0 era, Google rose. In the Web 2.0 era, Meta rose.

In this AI-driven wave, where will the next Google or Meta come from?

I'm not sure. But here's what I feel fairly confident about:

The platforms (Google, Meta, TikTok) have users, data, and a first-mover advantage. But they have to keep innovating — the moment the channel shifts, that advantage starts to dilute.

For tool-building startups, the window is these next few years. Whoever can crack the hard problems — "cross-platform ad buying," "influencer marketing automation," "landing pages for the agent era" — will earn a seat at a brand-new table.

For brand builders, you need to accept one thing: the audience you're winning over is migrating from "humans" to "humans + agents." What you do now to prepare will determine whether you can still advertise effectively two years from now.

I don't know who wins in the end.

But there's one thing I'm fairly certain of: brands that still treat AI as just "a way to save a few headcount" will likely get cleared out in this reshuffle like last season's inventory.

That's my judgment — it may not be right. But it's a question worth thinking about right now.