The stakes for the average internet user are getting higher. Ask a chatbot a question — about a drug interaction, a flight delay, a product recall — and the answer might sound confident. It might even be wrong.
A startup called Seltz has raised $12.5 million in seed funding to build a search engine that tries to stop that from happening. But the company is not building another chatbot.
It is building the infrastructure underneath them. The money comes from a group of investors who see the search wars flaring up again. The source material describes the current moment as a rekindling of a long-dormant fight.
For years, web search was a settled game. Google won.
Everyone else played catch-up. Then AI agents showed up. These are the software bots that book your travel, compare your products, check your medical research.
They need information fast. They need it to be true. And the old search architecture was not built for them.
Seltz wants to fix that at the core. Not by polishing the user interface.
Not by adding a chat window. By rebuilding how machines find information. The company is focused on surfacing timely, relevant content about news, products, and services — but in a format that AI systems can actually trust.
For a patient looking up a new treatment, or a parent trying to get a quick answer, the difference matters. The current generation of large language models struggles to separate fact from plausible-sounding fiction.
Seltz is building the layer underneath that prevents those errors from happening in the first place. This is not a solo effort. The source material notes that Seltz faces competition from other startups and established players who are also chasing the same prize.
The race is on. And the prize is big. Whoever builds the trusted search layer for AI agents effectively controls how machines see the web.
That is a different kind of search monopoly than what we have now. It is not about serving ads to humans.
It is about serving facts to bots. The human side of this is easy to miss. The technology is abstract.
The funding rounds are big numbers. But the concrete consequence is simple.
If Seltz succeeds, the chatbot that tells you a drug is safe will actually be right. If it fails, the hallucinations continue. The confident wrong answers keep coming.
The current experience is fractured. You ask a question. You get a plausible lie.
You have to fact-check the machine. That is not a feature.
That is a broken pipeline. Seltz is betting that the way machines find information needs to change at its core. The seed round gives them the runway to try.
The investors are betting that the search wars are real this time. That the old guard can be displaced.
That a startup can build the infrastructure for a new generation of information retrieval. For families, for patients, for anyone who relies on getting accurate information fast, the outcome of that bet will matter. It will determine whether the answers you get are true or just convincing.
