The Hair-on-Fire-Problem created by ChatGPT: How do marketers reach an audience that is leaving Google behind?

For twenty years, Google has dominated search. But, as OpenAI crosses the 400 million monthly active user mark, it’s clear that Google’s iron grip on search is loosening. And quickly. In my own experience, I’ve gone from an all-Google search history to a 70/30 split, with ChatGPT gaining more and more mind (or search) share every day. I expect that by the end of the year, over half my questions will go to AI-search vs. traditional search. 

Younger generations are moving away from Google even faster. I recently spoke with a class of students at Northwestern and used that as an opportunity to gauge their own search patterns. Nearly every student raised their hand when I asked them if they were using the new models in their daily lives to search for new products and services. One student said that for her, high school was the Google era and college had become the ChatGPT era. 

As goes search, so goes brand advertising dollars.

As eye balls move to AI-search engines like ChatGPT, Perplexity and DeepSeek, marketers are forced to grapple with a new reality – how do I reach an audience that is moving away from Google? Google birthed massive industries – Search Engine Optimization (SEO) and Search Engine Marketing (SEM) – for brands to connect with their customers. Digital marketers poured massive amounts of dollars into these online channels to optimize customer acquisition. In 2023 alone, brands spent close to $90B on branded and organic search.  

When we talked to marketers to understand this trend, we were surprised by the overwhelming interest in the topic. Mercury CEO Immad Akhund’s recent tweet is a good illustration of the work being done inside of companies to understand their AI search positioning:

The opportunity for founders

We initially see 3 areas that are ripe for builders to tackle in this new landscape: AI SEO, AI Advertising Networks, and AI Ad Generation. 

AI SEO: The most pressing topic we hear about is brand performance inside of the AI engines. How often does Starbucks appear in ChatGPT or DeepSeek when users ask questions about coffee? How do their results compare with Google? Profound has made a big splash in the category with some impressive customers like Rho, Rippling and Indeed. It’s clear this market is forming as we've seen a number of existing teams building as well – Quno, AthenaHQ, Bluefish – and expect AI SEO to be an established category CMOs talk about by the end of 2025. In this category, the Holy Grail product that (as far as we can tell) has not been built is a tool that helps you improve your rankings vs. just giving you visibility into how your brand is performing. OpenAI is the obvious company to build this product, but the value of a platform-agnostic tool will increase significantly with competition. Synthetic research may be a new tool for cracking SEO rankings in AI search, as my colleague Bohan has already explored. Whoever manages to build a great product here will meet a flood of inbound interest and $$$ from digital marketers.

AI Advertising Network: The shift to AI-driven search opens a massive opportunity for an AI-native ad network, enabling brands to place targeted, context-aware ads within AI-generated answers and conversations. Just as The Trade Desk ($30B) reshaped programmatic advertising and AppLovin ($90B) built a dominant mobile ad network, a first mover in AI-native advertising could capture billions in ad spend as brands seek visibility in AI-driven search and recommendations. It’s worth noting that neither of these legacy juggernauts were obvious but built huge outcomes and defensibility on top of advertising platform shifts. Let us (meaning us as a VC ecosystem) not repeat the same mistake of overlooking this time. Nexad and Koah are teams starting to build in this direction and we wouldn’t be surprised if more startups chase this theme as well.

AI Ad Generation: As the medium for discovery shifts, so too will the content. The cost of advertising production is heading to zero – a trend our friend Rex Woodbury covered well in a recent post on Digital Native. Advertisers today are using AI models and creative tools to generate prototype ad images (Midjourney, Leonardo, Krea), ad videos (Runway, Pika), and ad copy (Jasper).  We believe there will be AI ad generation specialists that not only generate the creative, but help you personalize and target your audience. Companies like Icon have already made it easy to create ads in minutes that would have required thousands of dollars and production teams in the past. Creatify.ai is building an early following as well. And Tofu is already bringing hyper personalization to the enterprise world. 

In the (near) future, we believe the AI search platforms will develop their own versions of ad networks similar to Google Adsense. Perplexity has been the most public in their intentions to monetize via ads but there’s no doubt the other AI search engines are scheming on their own advertising models. As Ben Thompson recently wrote in Stratechery — “the only way for a consumer tech company to truly scale to the entire world is by having an ad model.” However it develops, the toolkit for ad generators will become increasingly important.

The naysayers will argue that AI adtech is false gold for founders. 

And they may be correct. Nobody knows how ChatGPT, Perplexity, Gemini and others will think about advertising; perhaps they’ll take Walled Garden approaches that make it hard for third parties to thrive. If ChatGPT’s consumer monopoly holds it may make these tools less interesting, but then again search and social were generally monopoly markets (Google/Meta) and still managed to enable large platforms like Trade Desk and AppLovin. Finally, it’s easy to see a world where some companies take an early lead only to get rug-pulled in the coming years. 

And, of course, advertising/marketing tech has become a graveyard for countless companies who tried to build around previous technology shifts like mobile and social and failed. How many logos from this dizzying market map in 2018 are relevant today? Not many. 

Ignore the haters

What’s clear at this point is that what worked for brands in the pre-ChatGPT era of 2022 will not work for brands in the coming years. There’s a Salesforce-sized opportunity for whoever figures out the tools that shepherd in a new era of digital marketing. If you’re building in this space or have a perspective on this trend, we’d love to hear from you.