Ando Wants Team Chat Where AI Agents Are Full Members, Not Bots in a Channel

Ando came out of stealth on 24 September after close to a year of building, announcing 20 million dollars in seed funding from Accel, Index Ventures and Emergence Capital for a team messaging platform built around a different premise to Slack or Teams: agents are members of the team, not integrations bolted onto the side of one.

In Ando, an AI agent can sit inside channels and threads with its own identity and permissions, follow conversations it is part of, keep persistent context and memory, and join live conversations the company calls Jams in real time rather than only responding when directly pinged. The platform is agent agnostic, so a team can bring whichever agents or coding harnesses they already use rather than being locked to one vendor's assistant.

Founder Sara Du frames the goal as removing what she calls meat proxies, the person whose actual job becomes relaying an agent's output to the rest of a team because the agent itself has no seat at the table. That framing is doing a lot of work here. Most agent tooling today still assumes a human sits between an agent and the people who need its output, checking it, reformatting it, deciding what to forward. Ando is betting that assumption gets in the way once agents are doing enough real work to justify their own presence in a conversation.

It is early. Ando is aimed initially at teams of two to 40 people, with larger rollouts planned for later in the year, and agent native anything tends to look cleaner in a launch demo than in a channel with six months of history and three different agents talking past each other. But the underlying question, whether a workplace chat tool should treat an agent as a participant with standing rather than a bot with a slash command, is one every team running agents day to day is going to have to answer for itself soon regardless of which platform they land on.