Instinct and Caddy lead surge of AI agents in text messaging

This survey of text based AI agents shows how quickly agentic AI is moving from developer playgrounds into everyday communication channels. Instead of asking users to install yet another app, these services ride on top of iMessage, RCS and other chat platforms, turning phone numbers into persistent agents that remember, schedule and transact. That is a subtle but important shift: the user interface for early agent systems is converging on the existing messaging fabric rather than on bespoke clients or headsets.

In the race to AGI, consumer adoption curves matter because they determine where real world data, feedback and money flow. If companies like Instinct and Caddy can prove that messaging native agents deliver repeat value without feeling creepy, they will create pressure on big platforms to embed comparable capabilities at the OS level. That in turn will increase demand for models that are not just good at chat, but that can reason over long lived personal context and call tools safely.

Competitive dynamics here are asymmetric. Startups can innovate on UX and niche workflows faster, but Apple, Google and Meta ultimately control the rails. If platform owners decide they want to be the primary provider of agents inside Messages, RCS or WhatsApp, many of today’s independent agents could be squeezed into narrow verticals or acquired. For now, the proliferation of options gives the ecosystem valuable experiments in what everyday users actually want from an AI that lives in their chats.