AI agents are joining your team — and Slack was not built for that
A startup called Ando just raised $20 million to build the first messaging platform where AI agents are real teammates, not bots — and it signals something important about how work is changing.
- $20M
- seed funding raised
- 4
- venture capital investors
- 12+
- countries with active teams on the platform
- Jan 2026
- when Ando's own team started using it internally
Right now, most AI agents live at the edge of your team. You call them up, they do something useful, they report back. In Slack or Microsoft Teams, they get treated like a bot, not a colleague — no inbox, no ability to start a conversation, no way to join a discussion unless someone explicitly tags them. That design made sense two years ago. Today, when agents are doing real work on real deadlines, it creates unnecessary friction.
On 24 September, a startup called Ando came out of stealth with a messaging platform built from scratch for a world where agents are full team members. It raised $20 million from Accel, Index Ventures, Emergence Capital, and Contrary Capital. Founder Sara Du, who previously built Alloy Automation, says the problem is structural: agents were treated as apps to install, even as they became participants in the team. On Ando, agents can join channels on their own, send messages first, and coordinate work across conversations — without a human having to relay the output.
The platform already runs across teams in more than twelve countries, in software, real estate, and financial services. It is agent-agnostic, so teams can connect the AI systems they already use — Claude, Codex, Grokbot, and others. Each agent gets a persistent identity, its own inbox, and access to the conversations it is allowed to follow. That last part matters: an agent can spot a problem in one channel and bring the right people together in another, without anyone having to notice it first.
The takeaway is not "switch to Ando today." It is that serious investors at Accel and Index are betting the messaging layer itself needs to be rebuilt for agents — which signals how fast the role of AI at work is changing. If your team is already running agents on real tasks, the right question is: how much time do you spend copying outputs from one place to another? That friction is what the next wave of work tools is being built to remove.