Sara Du's solo venture Ando secures $20 million in funding
Sara Du previously co-founded Alloy Automation, a company focused on e-commerce automation and integration, which received support from Y Combinator and funding from investors including Andreessen Horowitz. She is now launching Ando as a solo founder, positioning it as a team messaging platform for humans and AI agents.
Ando publicly launched from stealth mode on September 24, 2026, announcing a total of $20 million in pre-seed and seed funding. Accel led the pre-seed round, while Index Ventures and Emergence Capital led the seed round. The product treats AI agents as full team members, granting them independent identities, permissions, and context, allowing them to join channels, actively participate in conversations, and initiate direct messages.
The platform supports agent-agnostic access, compatible with external agents like Claude, Codex, and Devin, while also providing its own hosted native agents. It currently focuses on small teams of 2 to 40 people, charging based on human seats rather than measuring agent actions. Early users are spread across about 15 countries, covering sectors such as software, real estate, and financial services.
Du stated in the Solo Founders podcast that most markets have only a few serious players, while others focus more on short-term funding rather than long-term issues. She believes agents should be treated as colleagues rather than tools and pointed out the structural limitations of existing platforms like Slack in the AI era. The Ando team has been working solely on the product internally since the beginning of the year.
Following the funding, the company plans to expand the platform and hire more staff. Du previously served as CEO of Alloy for about five and a half years and has now chosen the solo model to advance Ando, emphasizing that the problems need a ten-year investment. The product includes channels, direct messages, group chats, and transcribable real-time calls, with agents able to monitor channels and respond without needing to be @mentioned.
In terms of market mechanics, investors are betting that agent collaboration requires independent infrastructure rather than layering on existing tools. The funding is directed towards a coordination layer designed specifically for human-agent hybrid teams. The beneficiaries are small teams that adopt agents early, as traditional messaging platforms struggle with the native experience of agents. The charging model prioritizes locking in payments from human users, lowering the barrier for agent usage.
Source: Public information
ABAB AI Insight
Sara Du founded Alloy Automation at the age of 19, starting with vertical automation in e-commerce and later expanding into a general integration platform, supported by institutions like a16z. She was a Thiel Fellow and focused on product and design after dropping out of Harvard. During her assistance in building MCP servers for enterprises in 2025, she observed the access barriers for agents in tools like Slack, leading her to initiate Ando.
On the capital path, the $20 million injection from Accel, Index, and Emergence is intended to build a natively designed messaging layer for agents from scratch. The funding supports product development and early team expansion while maintaining the solo founder structure. Du's choice of the solo model is to align with her strengths in defining long-term problems and product taste, avoiding dilution of direction by co-founders.
This is similar to how Slack evolved from an internal tool to a standard collaboration layer, or how Notion expanded from a note-taking tool to a work operating system. Current collaboration software is transitioning from a human-centric phase to a human-agent hybrid stage, where agents need independent identities and permissions rather than being integrated plugins. Ando aims to define this category.
Essentially, this is a reconfiguration of the interface layer in the restructuring of the industry chain. Old platforms assume participants are human, with messages and context designed around humans; the continuous memory, proactive actions, and multi-tool permissions of agents require a new coordination medium. The mechanism is that when agents become everyday executors, the cost of contextual fragmentation increases, and those who provide a unified participation interface can capture decision-making and task flows, forming a new entry point for work operating systems.
ABAB News · Cognitive Laws
Tools become colleagues, interfaces must be rewritten
Most competitors are just noise
Only ten-year problems are worth solo bets