The tab you already have open
The average smartphone user checks their messaging app 60-80 times a day. It's the single highest-frequency app on most people's phones. Any new app, by definition, is competing for fewer daily opens than Messages already has.
If the whole point of personal AI is to be present in your life, the interface question answers itself. You don't build a new app and ask users to switch to it. You meet them in the app they're already in.
What's special about iMessage specifically
iMessage isn't just a messaging app. It's deeply integrated into the OS — it sits in Notification Center, in Siri, in the Messages widget, in CarPlay, on the Apple Watch, on the Mac. A contact in iMessage is available at every interaction layer of an iPhone user's life.
No third-party app gets that real estate. You can't build your app into Notification Center. But when your AI is a contact in iMessage, you already have it.
The zero-install benefit
Personal AI fails conversion on installation more than any other step. Every install screen is a cliff where 40-60% of interested users drop off. The iMessage-native pattern skips the cliff entirely — you text a number, you're in.
The TTFV (time-to-first-value) drops from minutes to seconds. That's why iMessage-native products convert so much better at the top of funnel.
Trade-offs, honestly
iMessage isn't free. Some real constraints:
- You can't render custom UI. No rich widgets, no interactive components. Text (and iMessage's native reactions/tapbacks) only.
- SMS fallback for Android means a narrower feature set for a chunk of your audience.
- Delivery isn't guaranteed — carrier-level issues exist.
- You rely on third-party infrastructure (Sendblue, Twilio) to actually send/receive.
- No direct push — you can't send a notification outside the messaging thread.
For an AI agent that needs to book flights with a visual confirmation, iMessage is the wrong interface. For a companion whose job is conversation and memory, it's the right one.
Why it's underbuilt
Given how clearly iMessage-native is the right surface for personal AI, why aren't more startups there? Three reasons:
1. Getting a paid iMessage line is non-trivial
You need a provider like Sendblue to register a paid iMessage business line. That's not the same as sending SMS over Twilio. A lot of developers don't know it exists.
2. App-store narrative
Founders default to 'we're building an app' because that's the mental model for consumer. Pitching 'we're building a contact' confuses some investors. It shouldn't — the unit economics are often better — but it does.
3. Feature-flex resistance
You can't show off a beautiful UI if there is no UI. Teams with heavy design DNA sometimes resist the plain-text constraint.
Where it's going
More consumer AI will land in messaging. Poke raised $100M building across iMessage/SMS/Telegram/WhatsApp. Brumo focuses on iMessage-first. Others will follow. The interface is obvious; the question is who executes.
Apple Intelligence in iMessage is another variable. If Apple builds a first-party AI into the Messages app, it could compress the category. But Apple Intelligence so far is tool-shaped, not companion-shaped — it answers questions rather than holding a long relationship. That niche stays open.
The bottom line
If you're building personal AI that's meant to live in someone's daily life — not a tool for discrete tasks — iMessage is the best interface available in 2026. The user is already there. The zero-install conversion is massive. The OS-level integration is free. The plain-text constraint is actually a feature: it forces you to make the conversation good, not the UI busy.
That's the bet Brumo is built on.