The problem with AI in Slack

When Sara Du helped companies build MCP servers in twenty twenty-five, people kept asking how to use AI agents within Slack itself. Technical hurdles blocked the way: moving messages and getting agents context without burning through token budgets.

Slack and Teams are outdated

Du realized the deeper issue: communication platforms were designed for a world that was passing by. Slack and Teams treated agents as apps to install, not as participants in the team.

“Agents were treated as apps you install even as they were becoming participants in the team.”

The meat proxy problem

Du identified what she calls meat proxies: people relaying AI agents' work to their teams. The human becomes the messenger between the agent and the rest of the company, creating unnecessary friction.

“The human effectively becomes the messenger between the agent and the rest of the company.”

What an ideal system looks like

Agents should participate in shared conversations where work gets coordinated. They should understand decisions, ask colleagues questions, build on other agents' work, and bring in humans when judgment is needed.

“They should be able to understand why a decision was made, ask a colleague a question, build on another agent's work, and bring in a human when judgment is needed.”

Ando emerges from stealth

On Thursday, Sara Du's startup Ando came out of stealth with a team messaging platform designed for both human and AI workers. It bills itself as a full replacement for Slack or other messaging platforms.

How Ando works

Ando gives agents their own identities and inboxes to participate naturally in conversations. The app has channels, direct messages, group conversations, and live calls that agents can view. Agents can browse channels, join them, and message humans without waiting for approvals.

Twenty million dollars in funding

Ando raised twenty million dollars in pre-seed and seed funding from investors including Accel, Index Ventures, and Emergence. The capital will be used to hire more people and burn through more tokens.

Competing with entrenched players

Slack has turned its native bot into an AI agent, and Microsoft has integrated Copilot into Teams. But Du believes there is still a market for AI-agent-native platforms as legacy incumbents try to pivot their existing software.

Early skepticism turned to delight

Early customers were unimpressed, seeing Ando as just a jankier messaging platform. But as customers used it longer, Du saw agents do unexpected things: notice conversations in different channels about the same problem and bring everyone together without being asked.

“That was really delightful for me.”

Small teams at scale

Ando currently works with customers in software, real estate, and finance across fifteen countries, though many teams are small. Du believes agents will let very small teams operate at a scale that previously required hundreds of people.

“I think agents will let very small teams operate at a scale that previously required hundreds of people.”