Google's Wednesday update to Gemini Live added new voice features and, in the process, made plain a design problem the company has been building toward for months: the Gemini app now contains three separately branded features — chat, Spark, and Daily Brief — each with its own icon and its own place in the app's navigation. Google's own announcement acknowledges the confusion in the very act of trying to dispel it.
The line from Google reads: "You shouldn't have to guess whether a task requires Spark, a Daily Brief, or a quick inbox search." That is meant as a promise about voice handling. Read another way, it is a confession that the app has grown three product surfaces that users are now expected to keep straight.
Daily Brief is the clearest example of engineering thinking leaking into consumer product. It is an AI-generated agenda that pulls from Gmail and Calendar to deliver what Google calls "proactive, personalized updates." In practice, the feature struggles to distinguish an urgent calendar conflict from an unprompted nudge to continue a chatbot conversation from last week — or, more awkwardly, from a resurfaced prior Google search. A user who was researching college scholarships or animal rescues does not necessarily want an AI raising the topic days later.
Key facts
- 01Google's Gemini app now splits into three separately branded surfaces: chat, Spark, and Daily Brief, each with its own icon and navigation slot.
- 02Daily Brief pulls from Gmail and Calendar to deliver 'proactive, personalized updates,' including resurfaced prior Google searches.
- 03Anthropic's Claude splits into 'Chat' and 'Cowork' modes, which until this week did not share memory of past conversations.
- 04ChatGPT similarly requires users to swap between 'Chat' and 'Work' modes.
- 05Apple's Siri approach embeds AI into existing surfaces — Spotlight Search, Photos, Camera — without asking users to learn new brands.
Spark has the opposite problem. It is arguably the more useful feature — an agent that takes action on the user's behalf — but Google has walled it off behind its own brand, its own tab, and its own mental model. A mainstream user should not need to decide which side of the Gemini app to be in before typing a request. The AI should route the task and spin up an agent when the task calls for one.
The pattern is not unique to Google. Claude currently splits into a "Chat" mode and a "Cowork" mode, and until this week, those two modes inside the Claude app did not share memory of past conversations — a state Anthropic only just corrected. ChatGPT runs a comparable split between "Chat" and "Work." In each case, the company is exposing its internal architecture — its inference surfaces, its agent runtimes, its memory scopes — directly to the consumer as separately branded products.
This is engineering-org design, not consumer design. Users are being asked to memorize brand names for what are essentially interaction modes powered by the same underlying model. It creates friction in a category that is supposed to reduce friction, and it competes for attention with the actual feature the user came to use.
The contrast with Apple is instructive. Apple's Siri overhaul has been criticized as anticlimactic, but the underlying approach — make Spotlight Search, the Photos app, the iPhone Camera, and voice requests quietly smarter without renaming any of them — is coherent. iPhone owners don't have to learn a new surface to benefit. The AI shows up where the behavior already lives.
The same logic is driving the emergence of text-based AI services — Poke, Ollie, Lindy, Orchid, Lucas, Folk, Tomo, and Instinct among them — where the user simply texts a chatbot and the assistant does what was asked. Text messaging is a universally understood interface. It doesn't require the user to figure out which product-within-a-product to open.
a16z investment partner Justine Moore recently framed the same idea: "People don't want to open an app every time they need help – they want a contact they can text like a friend. And the gold standard is iMessage." The bet behind the text-first crop of startups is that iMessage-style delivery beats a launcher screen full of AI sub-brands.
The counterargument for Google's approach is real. Spark and Daily Brief are distinct enough as capabilities that giving them names may help early adopters and reviewers reason about what the product does. Internal teams gain identity, roadmaps become legible, and the feature marketing has something to point at. That is how most software companies have historically built product lines, and it works — up to a point.
The point where it stops working is when a single app forces the user to choose between three branded surfaces before typing the first message. Gemini has arrived at that point, and Claude and ChatGPT are close behind. The next competitive edge in consumer AI is likely to come from whichever product hides the most of its own architecture. Apple has a structural advantage there because it can embed the AI in surfaces users already open by reflex. Google, Anthropic, and OpenAI will have to earn the same simplicity by collapsing brands rather than launching more of them — and the company that does it first will likely find the killer feature is not a feature at all, but the absence of a menu.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
