Ando has emerged from stealth with a $20 million pre-seed and seed financing package and an ambitious idea for office communication: give AI agents a place in the same conversations as their human colleagues. The company describes its service as a potential replacement for established team messaging tools, rather than another agent added to an existing chat app.
Founder Sara Du told TechCrunch that the concept grew from requests she heard while helping companies build MCP servers in 2025. Teams wanted to use agents through Slack, but moving messages, supplying context and controlling token use created friction. Her response was to design communication around agents as participants from the outset. The financing came from investors including Accel, Index Ventures and Emergence, according to the company’s announcement.
Agents get an identity inside the conversation
Ando gives agents their own identities and inboxes. The service includes channels, direct messages, group conversations and live calls that can be transcribed for an agent to review. Du says agents can browse channels, decide which conversations to join and contribute without being explicitly tagged. They can also notify a person when they decide human attention is needed.
The design tries to reduce a task Du calls the human relay: a worker repeatedly carries information between an agent and the rest of the team. In her proposed workflow, the agent can see why a decision was made, ask for missing context, share findings and involve a person when judgment is required. That is the company’s vision; TechCrunch does not independently measure how often the system performs those actions successfully or how much time it saves.
Some of those abilities also raise practical governance questions. An agent that can enter channels without being invited and send messages on its own needs clear boundaries around access and interruption. The source article describes Ando’s functions and the founder’s examples, but it does not detail permission controls, security testing or an independent assessment of agent behavior. Those omissions limit what can be concluded about how the product would operate in a large organization.
Funding arrives in a crowded market
Ando is not alone in adding agents to workplace communication. TechCrunch notes that Slack has developed its native bot into an AI agent and Microsoft has integrated Copilot into Teams. It also points to Buzz, a newer messaging product aimed more closely at developers. Ando’s wager is that a service designed with agents in mind can offer a different workflow from tools that added agent features later.
Du acknowledges that the early product did not impress every prospective user. Some saw a less polished messaging platform before they saw the intended advantage. She told TechCrunch that longer customer use gave the team confidence, citing an example in which an agent noticed related discussions in separate channels and brought the people together with context and a proposed decision. The example illustrates the desired behavior, not a published performance benchmark.
The company says it now works with customers in software, real estate and finance across 15 countries. Many of those teams are still small. Ando plans to spend the new capital on hiring and the computing costs involved in running agents, Du said. The article does not give customer counts, revenue or a breakdown of the two funding rounds.
The test is coordination, not chat alone
Du argues that small teams may eventually execute work at a scale once associated with much larger organizations if agents handle research, execution and coordination. That is a projection from the founder, not an observed result across Ando’s customer base. What the funding establishes today is investor support for experimenting with a different interface between people and autonomous systems.
The company’s next test is whether its agent participation remains useful when conversations multiply, permissions grow complicated and users already have established habits in rival platforms. TechCrunch’s reporting supports the existence of the funding, the product launch and early customer use. It leaves the operational benefits and the economics of running agents to be demonstrated over time.
