Sila

doing things differently co-founder/ceo @sila_hq (yc w26)

Sila is an agent-native messaging platform where teams and agents collaborate in the same conversations, with every agent working from the live context of everything the team says . It targets fast-moving companies whose day-to-day work now runs on a mix of human chat and AI output, and it treats agents as participants inside channels, threads, and DMs rather than something you invoke through a sidebar or slash command. The launch lands as the team makes the case that the communication tools most companies rely on are over a decade old, designed for a world that no longer exists, while 60% of employees say those tools make them feel overwhelmed , and AI has changed how work gets done by pushing up the volume of communication, the pace, and the coordination burden across an organization . In practice, that means channels, threads, and real-time messaging that look familiar on the surface, plus shared AI sessions that anyone in a workspace can see and contribute to instead of the private one-to-one chats most people run with ChatGPT or Claude today. Sila also supports a "bring your own agents" model, with an agents page and public docs aimed at connecting external agents into the same conversations where teammates already work. The pitch behind the launch tweet, that AI today is single-player while real work is not, extends into the product as team-agent and agent-agent interaction being first-class rather than bolted on. The company is Sila Intelligence, founded in 2026 by CEO Mith Paresh Patel and CTO Carl Huang out of New York City and is part of the Y Combinator Winter 2026 batch. For founders and operators evaluating where AI actually fits in their stack, Sila is worth a look as a bet that the next default workplace chat tool will be one designed around agents from the first commit rather than one that adds them later.

Tags

Comments (11)