This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
Alphabet reported its first $100 billion quarter yesterday. Revenue was $102.3 billion, up 16 percent, and Google Search & other grew 15 percent to $56.6 billion, according to the earnings release. The figure that should interest anyone whose company gets described in Google came in Sundar Pichai’s prepared remarks: AI Mode “now has over 75 million daily active users.”
Earlier calls this year emphasized monthly users of AI Overviews. Monthly users measure reach. Daily users measure habit. Pichai paired the AI Mode figure with two others: U.S. AI Mode queries doubled over the quarter, and Google “shipped over 100 improvements to the product in Q3.”
Investors will hear those three numbers as momentum. Issuers should hear them as a change in how they get looked up.
What a daily habit does
A daily user of a conversational search mode is not looking something up once. They ask follow-ups, refine, compare. When the subject is a company, the description arrives in pieces across a session: what the company does, then why the stock moved, then how it compares with a competitor, then whether the lawsuit matters.
Each step can pull different sources. The first answer might lean on the company’s own pages and a reference entry. The fourth might lean on a forum thread or a two-year-old article. Habitual users ask more of the fourth kind of question, because habit is what makes people comfortable asking a search box things they once would have asked a person.
Pichai also said the query-growth effect of AI Overviews was “more pronounced with younger people.” For financial firms thinking about the next generation of retail investors and clients, that line may matter more than the headline totals.
He added that AI Mode rolled out across 40 languages during the quarter. In 2023 I wrote that when generative answers go global, local reputation becomes a retrieval problem. The habit is now forming in many languages at once, often from thinner local sources.
What 100 improvements means
On an earnings call, 100 improvements in a quarter is a statement about speed of execution. For a company being described, it is roughly one change a day in the system producing the description. Some changes will improve accuracy. Some will change which sources are pulled or how an answer is laid out. A test run in July is not a baseline for October.
This is not a criticism of Google. Fast iteration on a product used by tens of millions of people every day is what you would expect. It is a reason to stop treating an audit of AI answers as a one-time project.
Where finance and IR teams come in
Earnings week is the obvious stress test. After results, people will ask AI Mode what the company reported, why the shares moved and what management said. The answer will be assembled from the release, wire stories and whatever commentary gets indexed first, and then extended through follow-up questions nobody on the IR team will see.
Three habits are worth adopting. Run your own earnings questions in AI Mode the day after results and again a week later, and keep dated transcripts. Ask the follow-ups an analyst or a skeptical shareholder would ask, not only the headline question. And make the plain-language version of your results easy to find on your investor pages, in sentences and not only in tables and slides.
Alphabet’s shareholders heard that generative search is growing the business, and on the numbers given it is. The issuer’s version of the same news is that more people now meet companies in an ongoing conversation, and the system running that conversation changes underneath it almost daily.