The Post-Slack Era: Why VCs Are Betting $20M on AI Coworkers

SAN FRANCISCO, CA – September 24, 2026 — Every decade or so, the fundamental architecture of the workplace undergoes a seismic restructuring. The filing cabinet gave way to the shared server; the cubicle yielded to the open floor plan; the lengthy email chain was usurped by the synchronous pulse of enterprise chat. Today, we are standing at the precipice of another structural rupture, but this time, the shift is not just about how we communicate. It is about who we are communicating with.

Ando, a stealth startup that has spent the last year quietly building an entirely new category of coordination software, emerged from the shadows today armed with a $20 million seed funding round. Backed by a formidable syndicate of venture capital heavyweights—Accel, Index Ventures, Emergence Capital, and Contrary Capital—the company is placing a massive bet on a provocative premise: legacy messaging platforms like Slack and Microsoft Teams are fundamentally ill-equipped for the future of work.

Instead of treating artificial intelligence as a bot to be summoned, Ando has designed an agent-native messaging platform from the ground up. In this new ecosystem, AI agents operate alongside humans as first-class members of the team, possessing persistent identities, granular permissions, and shared context.

Breaking the Legacy Chat Scaffolding

To understand the magnitude of the problem Ando is attempting to solve, one must look at the underlying architecture of current workplace communication. Platforms built between 2013 and 2016 were designed for human typists. Their core unit of value relies on human-to-human interaction, with bots retrofitted onto legacy frameworks as reactive add-ons, side-panel assistants, or simple slash-command triggers.

As organizations increasingly deploy autonomous AI models via frameworks like Anthropic's Model Context Protocol (MCP), a profound friction has emerged. Workers find themselves acting as human relays, endlessly copy-pasting outputs, errors, and instructions back and forth between external agent consoles and company chat channels.

“Every major computing shift creates new systems for how work gets coordinated,” said Ivan Zhou, Partner at Accel, who led Ando's initial pre-seed round. “Agents aren’t simply another feature inside existing workplace software. They’re becoming participants in companies themselves. Sara saw early that this shift would require an entirely new kind of workspace built around collaboration between people and agents. We believe that creates a new category of coordination software, and that Ando has the opportunity to define it.”

Ando eschews the walled-garden approach pushed by enterprise software giants. The platform is entirely agent-agnostic, allowing teams to plug in external models ranging from Claude and OpenAI Codex to xAI's Grokbot or custom local harnesses. By utilizing a three-layer technical stack—a human messaging layer, an agent execution layer, and a compounding memory layer—the software merges search, context, and composition into a unified action.

From Chatbot to Colleague

The cultural implications of treating software entities as equal colleagues are vast. In Ando, agents do not wait passively to be mentioned. They are autonomous, proactive participants that evaluate channel discussions and can chime in unprompted when they possess relevant data or solutions.

This dynamic is most vividly demonstrated in "Jams," Ando's real-time audio and video collaboration feature. Unlike legacy huddles where transcripts are passively attached as an afterthought, Jams are engineered with synchronous AI ingestion. All spoken dialogue is transcribed, indexed, and analyzed in real time.

Early adopters are already demonstrating how this transforms daily operations. At one mid-sized engineering firm currently beta-testing the platform, human developers utilize Jams for technical architecture debates. As they speak, coding agents listen to the live call, cross-reference the dialogue against existing issue histories and the codebase repository, and autonomously draft multi-page, implementation-ready specifications by the time the humans hang up. The typical hours spent on manual context assembly are entirely evaporated.

“This isn’t about whether Slack or Teams can add agents. Of course they can,” noted Sara Du, Ando's founder and CEO. “Starting from scratch, unencumbered by legacy scaffolding, lets us design around what agents are actually capable of rather than what existing software allows them to be. I think agents are capable of surprising us if we actually give them the permissions and the context to.”

Sara Du’s Second Act and the Architecture of Silence

Du is no stranger to the complexities of enterprise workflows. A 2022 Thiel Fellow who left Harvard University after her freshman year, she previously founded Alloy Automation, an enterprise integration platform that raised $27 million and scaled to serve major global brands. Her transition from building API connective tissue to reimagining the corporate organizational chart stems from a deep understanding of systemic bottlenecks.

Interestingly, the platform's design philosophy borrows heavily from its namesake, the Pritzker Prize-winning Japanese architect Tadao Ando. Known for his mastery of Ma—the concept of negative space or emptiness—the messaging software integrates silence into its user experience. To combat the notification fatigue that plagues modern workers, Ando's reasoning engine evaluates incoming queries. Before pinging a human coworker, the system checks if an agent or the company's compounding memory layer can resolve the issue. Notifications are collapsed into contextual digests, restoring focus to human workers who are increasingly overwhelmed by the digital noise of their own organizations.

The Governance of Autonomous Coworkers

Of course, granting autonomous software agents persistent identities and write-access within an enterprise environment introduces a labyrinth of security and governance challenges. The anxiety of prompt injections, runaway API bills, and unauthorized data access is palpable among modern IT leaders.

Ando addresses these realities through a strict, member-level boundary isolation framework. Agents inherit the exact same permission boundaries as their human counterparts; they cannot access private direct messages unless explicitly invited. Through a centralized control plane, workspace administrators can establish granular tool execution permissions—differentiating between an agent that can read a codebase and one that can deploy to production.

Furthermore, the platform establishes "per-channel token budgets" to prevent runaway iterative loops and spam generation, while immutable audit logs track exactly which human or agent initiated a specific action. It is a necessary bureaucratic scaffolding for a world where agents might soon outnumber human employees five-to-one.

The Connective Tissue for a New Workforce

The venture funding environment has definitively shifted from financing general foundation models toward building the orchestration layers that integrate these models into human team dynamics. With deployments already active across 15 countries in sectors spanning software, real estate, and quantitative finance, Ando is proving that the market is hungry for a native solution.

“The most important communication platforms have always grown into something bigger: the operating layer for how organizations work,” said Sofia Dolfe, Partner at Index Ventures. “Ando’s structured context model gives it the potential to connect communication, tasks and agent operations in a way legacy platforms can’t.”

This sentiment is echoed across the investor syndicate. “We’ve spent 20 years focused on productivity, and we believe the way people work with each other and with AI agents is going to change fundamentally,” added Santi Subotovsky, General Partner at Emergence Capital. “Teams need a platform that understands the context of their collaboration and helps them identify which tasks are better suited to people and which to AI.”

As artificial intelligence transitions from a novelty tool into a structural component of corporate labor, the software we use to coordinate that labor must inevitably evolve. By giving agents a permanent seat at the digital table, Ando is not merely launching a new messaging app; it is drafting the blueprint for the post-human-only corporation.