Ando raises $20M to replace Slack, while keeping Slack as a bridge

Why it matters

Ando is betting that agents need to join the conversations where teams make decisions, which means winning a messaging migration as well as an AI product evaluation. Its waitlist and small-team focus show how early that test remains.

Sara Du (@saradu)'s Ando emerged from stealth on September 24th with $20 million in pre-seed and seed financing, pitching a new team messaging platform where AI agents can join conversations as coworkers. In its report on the launch, TechCrunch named Accel, Index Ventures and Emergence Capital among the investors. Ando's launch materials also list Contrary Capital as a backer.

Du arrived at Ando after building a different kind of connective tissue for business software. She co-founded Alloy Automation in 2019, first making no-code tools for ecommerce merchants before shifting the company toward enterprise integrations. Before Alloy, she worked as an engineer at Snap and DoNotPay, according to Gold House's profile. She attended Harvard for about a year and took part in Dorm Room Fund before leaving to build Alloy, she said in a Dorm Room Fund interview.

The origin of Ando was similarly practical. In 2025, while helping companies build MCP servers, Du kept hearing a version of the same request: let people use agents directly inside Slack. Getting messages and context to the right agent without running up token costs exposed a larger product problem, she told TechCrunch. Existing chat platforms treat agents as add-on apps, Du argued, even as teams increasingly expect them to contribute to shared work. She described employees relaying agent output to colleagues as "meat proxies."

From chat integrations to a shared workspace

Ando's answer is to move the agent into the room where a team coordinates work. The product includes channels, direct messages, group conversations and live calls that agents can access through transcripts. Agents have their own identities and inboxes; they can browse channels, choose conversations to join and contribute without waiting for a person to tag them. They can also bring a human into a discussion when they judge that human input is needed, according to Du's description to TechCrunch.

That autonomy raises a question of access as much as convenience. Ando says agents cannot read direct messages unless a user forwards the context. Ando also says it is building for cloud and local agents, with APIs and developer tools intended to let teams use accumulated conversation context elsewhere. On its product site, Ando says it will maintain a Slack connection for teams during a transition, while calling Slack a bridge rather than the destination.

The pitch extends beyond faster bot responses. Du's bet is that an agent with access to ongoing conversations can connect work that people have split across channels, preserve the reasoning behind decisions and take on coordination tasks. She told TechCrunch that an Ando agent had spotted related discussions in separate channels and gathered participants into one group conversation with context and a suggested decision. Du offered this as an example; Ando has not disclosed how often the product does this or how reliably it works.

Ando has described early adoption in broad terms. Du told TechCrunch that Ando was being used by teams in software, real estate and finance across 15 countries. The launch materials describe teams in about a dozen countries. Ando says its own staff have worked exclusively in the product since January 2026. Neither account gives a customer count, paid-user figure or revenue measure, so the geography and use cases show reach, not commercial scale.

The expensive part is getting teams to move

Ando is challenging a product that already sits at the center of workplace communication. Slack and Microsoft Teams have added agent capabilities to their existing platforms, giving customers a way to experiment without moving conversations elsewhere. In July, Jack Dorsey's company Block launched Buzz, an open-source group chat product for teams and agents. The Buzz launch is another sign that the debate is shifting from whether agents can be bolted onto workplace software to where they should participate.

Ando's answer requires a bigger commitment than installing a bot. Teams must move their conversations, habits and, potentially, the context behind decisions into a new workspace. The Slack connection lowers the friction of starting, but Ando's own site makes clear that migration is the objective. The product is currently best suited, by Ando's account, to teams of up to 30 people plus agents, and Ando says it is still working through a waitlist. That puts its near-term opportunity among small teams that can switch tools without coordinating a company-wide migration.

Du has encountered the product's basic tradeoff firsthand. She told TechCrunch that early users sometimes saw Ando as a less polished messaging app and missed what was different about the agents. A chat product needs to be useful on its own before its agent model can matter; customers will not migrate just to validate a thesis about future work. Ando needs to make the agent participation visible and valuable enough to outweigh the inconvenience of moving.

Ando said the $20 million will fund hiring and the cost of running more agent activity. The financing is cumulative across pre-seed and seed rounds, rather than a single round size: Accel's portfolio page dates its initial investment to 2025. Ando did not attach a valuation to the announcement or break out the funding by round. The $20 million backs Du's thesis; it has not yet shown that teams are abandoning Slack.

Du has spent one company building software that connects disconnected business tools. With Ando, she is applying that instinct to the conversations that organize work, arguing that agents need access to shared context rather than a narrow channel into an employee's workflow. The test is whether that architecture makes coordination easier for real teams before the incumbent platforms make their own agent features good enough.