AI Coworkers Get a New Home as Ando Raises $20M
Sara Du just raised $20 million for Ando, a team messaging app built so AI agents can join the conversation as real teammates instead of bolted-on bots. The pre-seed and seed round came from Accel, Index Ventures, and Emergence, and it lands right as a lot of lean teams are already leaning on AI tools just to get through the week.
If you run a small team, this is not an abstract debate about whether an AI coworker belongs in your stack. It is about whether your project updates, client follow-ups, and status checks keep sliding through the cracks because a person has to manually relay everything an AI agent already knows.
Meat Proxies, Meet Your New Teammate
Du has a name for the problem she is trying to fix, describing the current setup as meat proxies. An agent finishes a piece of work, and a human has to carry that update into the group chat, the client email, or the weekly status meeting. Ando instead gives agents their own identity and inbox, so they can post in channels, join direct messages, sit in group threads, and even join live calls without someone relaying on their behalf.
The company already has customers spread across 15 countries in industries like software development, property management, and finance, though most of those teams are still small. Du has framed the industry’s default approach as bolting agents onto a workflow rather than welcoming them as participants, and closing that gap is the entire premise of the product.
Why Hybrid Teams Should Pay Attention
Most founders under 30 are not really choosing between remote and in-office anymore. They are managing a mix of employees, contractors, and AI tools that all need to stay in sync, and coordination is usually the part that breaks first. If you already run a chaos-free hybrid office, you know the hardest part is rarely the work itself. It is the handoffs between people and, increasingly, between people and software.
A messaging layer built to treat agents as participants instead of plug-ins could remove one layer of that coordination tax. That matters more for a five-person startup than for a 5,000-person enterprise, because a founder rarely has a chief of staff whose whole job is managing the relay.
What To Settle Before an Agent Gets a Seat at the Table
Before adopting any AI-native workplace tool, decide what an agent can say to a client or teammate without a human checking it first. Founders who skip this step tend to spend more time cleaning up mistakes than the tool ever saved them. The fix is not complicated. Borrow from the same delegation strategies for teams you would use with a new hire: define scope narrowly, set checkpoints, and review output on a schedule until trust is earned.
It also pays to track what these tools cost before they sprawl across every department. Founders who have already had to control ai spend across scattered logins and API keys know how fast a helpful pilot turns into an unmanaged budget line nobody owns.
The Incumbents Are Not Standing Still
Both of the category’s giants shipped their own agent features this year, and neither Slack nor Microsoft is short on engineers or cash. Du’s bet is that incumbents are constrained by architecture built around humans typing messages to each other, not agents reasoning through shared context. Whether that gap is real or temporary is the question investors just backed with $20 million.
| Detail | Figure |
|---|---|
| Funding raised | $20 million, pre-seed and seed combined |
| Lead investors | Accel, Index Ventures, Emergence |
| Countries with active customers | 15 |
| Founder | Sara Du |
According to McKinsey’s research on AI in the workplace, most companies are still early in redesigning workflows around AI rather than simply layering it onto old ones. That is precisely the gap a startup like Ando is trying to exploit before the giants catch up.
What to Watch as AI-Native Tools Multiply
Watch whether Ando’s early customers in those first few verticals stick around once their teams outgrow the size where a workaround still feels good enough. Watch too whether the incumbents respond with a genuine from-scratch redesign rather than another integration announced at a keynote.
For founders, the takeaway is not to switch tools tomorrow. It is to notice how much of your week already goes to relaying information an AI agent already has, and to start asking whether that job should belong to software instead of you.
Do AI agents really need their own inbox?
For teams already delegating real work to AI tools, yes. An inbox lets an agent receive direct requests and follow up without a human copying messages back and forth between systems.
Is this only useful for large companies?
No. Small teams often feel the coordination tax more acutely, since founders rarely have staff dedicated to relaying updates between people and tools.
