Google seeks to balance AI disruption with core business protection

Google has a problem most companies would kill for. It built some of the foundational technologies behind the generative AI revolution, and now it has to figure out how to use them without destroying the business that prints money.

That business is advertising, which brought in $237.8 billion in revenue. Deploying AI aggressively into search, the core product that drives those ad dollars, is a bit like performing open-heart surgery on yourself. You really want to get it right.

The innovator’s dilemma, Google edition

Clayton Christensen wrote the book on this exact scenario decades ago. A dominant company sees a disruptive technology coming, understands it intellectually, but moves too slowly because the new thing threatens the old cash cow. Google is living that playbook in real time.

Paul Buchheit, the creator of Gmail and one of Google’s earliest employees, has been particularly blunt about this dynamic. He’s argued that Google prioritized preserving its search monopoly over aggressive AI deployment. The result, in his telling, was institutional risk aversion that cost the company its edge in AI, opening the door for competitors like OpenAI and Microsoft to sprint ahead.

That criticism has some teeth. OpenAI’s ChatGPT went from zero to cultural phenomenon in weeks. Microsoft poured billions into the partnership and bolted AI into Bing, Office, and everything else it could find. Meanwhile, Google’s initial response felt halting. Its early Bard chatbot launched to lukewarm reviews and a factual error in its very first public demo. Not a great look for the company that essentially invented the transformer architecture powering all of these models.

A US antitrust ruling added another layer of complexity. The court found Google guilty of maintaining an illegal search monopoly. The ruling acknowledged that generative AI is likely to reshape search markets in fundamental ways, which means Google is simultaneously defending its dominance in court while trying to evolve the product that created that dominance.

How Google is threading the needle

Here’s the thing: Google hasn’t been standing still. The company has been rolling out its Gemini AI model across products and integrating what it calls “AI Overviews” directly into search results. These are the AI-generated summaries that appear at the top of search pages, attempting to answer queries before users even click on a link.

The delicate part is what happens to the ads. Google’s approach has been to weave generative AI into the search experience while carefully preserving ad placements. If AI answers everything directly and users stop clicking through to websites, the entire advertising model starts to wobble. So Google is essentially trying to make search smarter without making it so smart that nobody clicks on anything.

Beyond search, the company has deployed AI in ways that strengthen rather than cannibalize existing products. In 2024, Google used Gemini AI to remove more than 240 million policy-violating reviews and block 12 million fake business profiles on Google Maps. That’s the kind of AI application that makes an existing product better without threatening any revenue streams. It’s cleanup work, not disruption.

Google Cloud is another area where AI integration is more straightforward. Enterprise customers want AI tools, and Google can sell them through its cloud platform without the existential tension that exists in search. The competitive dynamics there are different: it’s Google versus Amazon Web Services and Microsoft Azure, and AI capabilities are a selling point, not a threat.

The SEO landscape is shifting under everyone’s feet

The ripple effects extend well beyond Google’s own balance sheet. The rise of AI-generated answers in search is fundamentally changing how businesses think about online visibility.

Traditional search engine optimization is evolving into what some are calling “LLM SEO,” which requires businesses to optimize their content not just for Google’s ranking algorithm but for conversational interactions with AI models. In English: instead of stuffing keywords into web pages, companies now need to think about whether an AI model would cite their content when answering a question.

That shift matters for every business that depends on organic search traffic, which is to say, nearly every business with a website. If Google’s AI Overviews answer a user’s question directly, the click-through to the underlying source may never happen. Publishers, e-commerce sites, and content creators are all watching this closely.

What this means for investors

Google’s parent company Alphabet is essentially making a massive bet that it can have it both ways: lead in AI while protecting advertising revenue. The early evidence is mixed.

On one hand, the company’s AI capabilities are genuinely impressive. Gemini is competitive with the best models available, and Google has advantages in data, distribution, and infrastructure that no startup can match. It has billions of users already interacting with its products daily, giving it a deployment surface that OpenAI can only dream about.

On the other hand, the structural tension is real. Every improvement to AI in search is a potential threat to the click-based advertising model. Google has to innovate fast enough to fend off OpenAI, Microsoft, and a growing field of AI-native search competitors, but carefully enough to avoid cratering the revenue engine that funds everything else.

The antitrust overhang adds uncertainty. If courts mandate changes to how Google operates its search business, the company’s ability to control how AI gets integrated into that product could be constrained. Regulatory risk and technological disruption hitting the same product at the same time is a combination that should keep investors attentive.

Look, Google has navigated major platform shifts before, from desktop to mobile, from links to apps, from text to video. Each time, skeptics predicted the end of its dominance, and each time, the advertising machine adapted. But generative AI represents something qualitatively different. It doesn’t just change how people find information. It changes whether they need to visit a website at all. That’s the question Google is trying to answer, ideally before someone else answers it for them.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.