Back 20 years ago when I spoke about search, I used to talk a lot about the “Google Habit.”  In the remarkably short time from its debut to the early 2000s, Google had become our defacto search choice. In fact, it was so dominant, we didn’t even consider its competitors. We didn’t think at all…we just Googled. That is how a habit works. We do things without thinking about them.

Fast forward 20 years. Google still dominates the information retrieval space. When we talk about worldwide search, Google delivers the results on 9 out of every 10 searches launched. It’s monolithic presence in the online landscape hasn’t really changed. But the way we navigate that landscape is beginning to. For the first time in a long time, Google has a real competitor when it comes to the way we look for our answers.

Another favorite topic of mine, following hard on the heels of the “Googe Habit,” was talking about the “usefulness of search.” I argued that while Google did a good job of retrieving information, it feel well short of the goal of making that information immediately useful to us. Today, with agentic A.I., the tantalizing promise of usefulness has finally arrived. The question is, what tools will we use to mine that usefulness?

The battle seems to be between Google’s Gemini, deeply embedded in the entire Google ecosystem, and OpenAI’s ChatGPT, a standalone app. Anthropic’s Claude is currently focused on the enterprise AI market.

Back two decades ago, I envisioned search gradually disappearing “under the hood” of various apps that made our lives easier. The act of actually retrieving information would be one step removed from us. What we would interact with would be a distillation of that information, formed into something we could use to do the things we wanted to do. What I didn’t anticipate was the emergence of the Large Language Model that currently powers ChatGPT and other AI models.

But as it currently stands, the act of retrieving information that we can use and the act of processing huge reams of text to predict useful responses are quickly converging. For an ever-increasing number of queries (currently about 50%) Google’s Gemini AI answers are now predominately displayed in the prime real estate of the search results page, straddling the very top of the “Golden Triangle.”  And ChatGPT is increasingly asked to retrieve specific information as users interact with its chatbot. Both Google and OpenAI realize that the future lies in a hybrid of the two.

The battle now is for either Google or OpenAI to dominate the various user interfaces in our personal technologies. And in this regard, it will be hard to beat Google. This January, Google and Apple inked a deal that would make Gemini the foundational AI platform for a future version of Siri. It’s already embedded in all Android devices. OpenAI’s early mover advantage over Google in terms of consumer AI usage is rapidly disappearing, with the two currently running almost neck and neck (36.6% for ChatGPT vs 27.4% for Gemini, according to a recent Emarketer forecast).

Brian X. Chen, lead consumer technology writer for the New York Times, indicated in a recent article that Emarketer’s numbers could be understating the competitiveness of OpenAI’s rival.  Google said at their recent Google I/O conference that in one year, the number of people using the Gemini chatbot more than doubled to 900 million. That puts it in a dead heat with ChatGPT and, if current growth rates continue, moving well past it in the next year.

Chen points out another massive advantage for Google’s Gemini over OpenAI’s ChatGPT. While OpenAI’s numbers are not public, they likely lost between $8 and $9 billion last year. Even the most optimistic forecasts don’t put OpenAI in a profitable position til 2029.

Google, thanks to its dominance in the online ad space, made a profit of $112 billion in 2025 on revenue of $386 billion. It will be relatively easy for Google to fold Gemini into that vast advertising supported ecosystem, moving to a cash positive position almost immediately. Because of the enormous development costs of AI, it’s hard to argue with the logic that player with the deepest pockets will be the ultimate winner.