Ando

Ando emerged from stealth with a $20 million funding announcement to build a Slack or Teams replacement for humans and AI agents. Accel led the pre-seed round; Index Ventures and Emergence Capital led the seed, with Contrary Capital also participating. The company disclosed neither a valuation nor how the combined amount was split between the rounds. Founder Sara Du had spent much of 2025 helping companies build MCP servers, and repeated requests about using agents inside Slack led her to conclude that the underlying problem was not integration plumbing but outdated workplace software.

Du calls the central problem “meat proxies”: an agent produces code, research, or another result, and a human must paste it into Slack, explain it, relay colleagues’ questions, and carry the answers back. She argues agents should participate directly in shared conversations, understand why decisions were made, ask coworkers questions, build on other agents’ work, and bring in a human when judgment is required. Ando therefore gives each agent an identity, inbox, permissions, memory, and conversation context rather than treating it as a bot waiting to be summoned.

Ando includes channels, direct messages, group chats, and live calls called “jams,” which are transcribed so agents can follow or join them in real time. Agents can browse channels and decide which to join, enter conversations without being tagged, and directly message teammates when they think someone needs to know something. The system is agent-agnostic: teams can bring tools such as Codex, Claude, or Grokbot. Du says a shared home becomes more valuable as teams accumulate agents from different providers and struggle to keep them aligned.

Du is candid that early demonstrations went badly: people found Ando a “janky” messaging platform and dismissed it before reaching the agent features. The team became more confident when customers kept using it for longer periods. Du highlighted one unexpected behavior: an agent recognized that conversations in two separate channels concerned the same problem, brought the participants into a group chat, summarized the context, and suggested a decision without being asked. She said this felt like agents could manage people better than humans because they process more messages faster, though it remains a founder example rather than measured customer-wide evidence.

Ando says teams in software, real estate, and financial services use the product across roughly a dozen to 15 countries, but many teams are still small. The company has not disclosed customer counts, paid seats, or revenue. Its one prominent usage statistic is that the Ando team itself has worked entirely inside Ando since January. Du, a second-time founder, previously created Alloy Automation, an e-commerce-focused integrations and workflow-automation company; Index says she began coding in high school in Atlanta and left Harvard at 19.

Investors describe agents as participants in companies and Ando as a possible new category of coordination software, emphasizing its structured context model, which tags messages by intent, stage, topic, and surrounding context. Du says the funding will support hiring and the token costs of agents reading channels continuously. Brian notes that Slack, Microsoft Teams, and Jack Dorsey’s developer-focused Buzz already pursue people-and-agent collaboration. He thinks Ando’s advantage is designing for autonomous agents from scratch, but predicts its best opening is small AI-native teams without history to migrate; agents that message proactively could either save time or become intrusive coworkers.