What Just Happened

On September 24, 2026, a team messaging startup called Ando came out of stealth with $20 million in seed funding from Accel, Index Ventures, and Emergence Capital. It's led by second-time founder Sara Du, who previously built the enterprise API platform Alloy Automation.

Ando's pitch is that AI agents aren't just apps or integrations bolted onto a chat tool, they're active participants in the team. Inside Ando, agents get their own identities and permissions, can follow channels, retain context across conversations, join live discussions called Jams, and message coworkers on their own initiative instead of waiting to be tagged.

Why This Is a Different Model, Not Just a New Chat App

  • Agents get identity and permissions, not just an API key. Treating an agent as a named participant with defined access, rather than a background process, changes how a team can reason about who did what and why.
  • Agents can initiate, not just respond. An agent that can join a channel and speak up unprompted is a genuinely different working relationship than a bot you have to remember to invoke.
  • It is agent-agnostic by design. Ando supports agents built on systems like Codex, Claude, Devin, and Grokbot, so a team is not locked into one vendor to get agents working alongside people.
Quick Insight

Ando is pricing per human seat, with agents themselves not charged per message or action, and it's initially targeting small teams of 2 to 40 people already running AI-native workflows. That is a signal the market for "AI as teammate," not just "AI as tool," is being built for small, fast-moving teams first.

What This Means If Your Team Already Uses AI Agents

Whether or not you adopt Ando specifically, the underlying shift is worth planning for: agents are moving from something a person invokes to something that has standing presence, permissions, and initiative inside your everyday tools. That raises real questions about access control, accountability for an agent's actions, and how a team decides what an agent is allowed to say or do without a human approving it first.

What We'd Tell a Client Right Now

Before giving any AI agent a standing seat in your team's daily workflow, define its permissions and its escalation path as deliberately as you would for a new hire: what it can see, what it can send without approval, and who is accountable when it gets something wrong. The technology for agents-as-teammates is arriving fast; the governance around it is what actually determines whether it helps or creates a mess.