Meet Your New AI Coworker: Skydive Wants to Put an Agent on Every Team

SAN FRANCISCO, CA – August 20, 2026 – The modern office is about to get more crowded, but not with new human hires. Anything, the AI app builder with over a million users, today launched Skydive, a platform that promises to fundamentally alter the composition of corporate teams. The premise is as simple as it is audacious: what if any employee, regardless of technical skill, could build and deploy their own AI coworker to take on a specific job?

This isn't another chatbot or a simple workflow automator. Skydive's creations are designed to be persistent, autonomous 'agents' that function like full-fledged members of the team. Each AI coworker is given a name, a role, and, crucially, its own dedicated cloud computer complete with a browser, file system, and terminal. They operate across a company's digital nervous system—Slack, email, iMessage, and the web—to carry tasks from request to completion, even when their human counterparts have logged off for the day.

Founded by former Google product leaders Dhruv Amin and Marcus Lowe and backed by a roster of high-profile investors including Bessemer Venture Partners and Shopify CEO Tobi Lütke, Anything is betting that the future of productivity lies not just in using AI, but in creating it. Skydive represents the company's bold second act, moving beyond building apps to building an entire digital workforce.

Beyond Automation: The Rise of the 'AI Coworker'

The term 'AI agent' has become ubiquitous, but Skydive aims to elevate the concept. Unlike traditional Robotic Process Automation (RPA) which mimics human clicks in rigid, predefined scripts, or workflow tools like Zapier that connect apps in a linear fashion, Skydive's agents are built for autonomy and adaptability. They are designed to understand context, make decisions, and pursue goals, not just follow a checklist.

This distinction is at the heart of the 'AI coworker' paradigm. These agents are not ephemeral tools summoned for a single task; they are persistent entities. The press release details several of these agents already at work within Anything's own 15-person team. 'Emma' handles customer support tickets in Intercom, a job she was built for in a single weekend by a support lead after an acquisition tripled ticket volume. 'Grace' investigates software bugs by accessing production logs and the company's codebase, allowing non-engineers to ship fixes. 'Rice' acts as a superhuman Sales Development Representative, researching leads and scheduling demos over email, while 'Justin' simply orders the office snacks via Instacart based on requests in a Slack channel.

These examples illustrate a system where AI is not just assisting with a task, but taking full ownership of a role. The agents can even collaborate. When Emma, the support agent, and Ellis, an agent who guides customers through app store submissions, need to coordinate, they communicate in a group chat, mirroring the behavior of their human colleagues. This level of integration and role-based specialization is what separates the 'AI coworker' from earlier generations of automation, marking a significant step towards a truly 'agentic enterprise' where humans and AI collaborate in complex, dynamic ways.

The Citizen Agent Builder: Democratizing AI in the Enterprise

Perhaps the most disruptive aspect of Skydive is its target user: everyone. The platform is engineered to empower the 'citizen agent builder'—the marketing manager, the operations lead, or the customer support specialist who understands a problem intimately but lacks the coding skills to solve it with traditional software.

"You don't need permission from IT to hire a coworker," the company's launch material states. "You just need to know the job, and then you can build the coworker who does it." This philosophy of radical decentralization challenges the conventional, top-down approach to enterprise AI, where development is centralized within specialized engineering teams. According to industry analysts, this centralized model often becomes a bottleneck, resulting in long development cycles and a limited number of high-impact agents.

Skydive flips the script. "Our mission has always been to make it extremely easy for regular people to build," said Dhruv Amin, co-founder of Anything. "Apps were the first version of that. Agents are next." The platform abstracts away the complex infrastructure—model access, sandboxing, security, and compute—allowing the user to focus on defining the job. Users can start from templates for common roles like 'Chief of Staff' or 'SEO Specialist' and then teach the agent the specific nuances of their company's tools and processes.

Anything has been its own most aggressive user, deploying over 150 agents to work alongside its 15 human employees. This internal dogfooding has allowed the small team to manage a product used by over a million people, a scale that would typically require a much larger organization. According to Amin, these agents "haven't replaced a single hire. They’ve taken on jobs we never had the headcount for and given every person here leverage that used to require an entire team.”

Reshaping the Modern Workforce and Organizational Structure

The implications of this technology extend far beyond individual productivity. If widely adopted, platforms like Skydive could fundamentally reshape organizational structures and the nature of work itself. The vision is one of an AI-augmented enterprise, where lean human teams can achieve unprecedented scale and efficiency by managing a workforce of specialized AI agents.

This inevitably raises questions about job displacement. While Amin emphasizes augmentation over replacement, the long-term economic impact of technology that can automate entire job functions remains a critical debate. The more immediate effect, however, may be a large-scale evolution of human roles. As agents take over repetitive and operational tasks, human workers will increasingly shift to roles centered on strategy, creativity, critical judgment, and, importantly, the management and direction of their new AI coworkers. The most valuable skill may no longer be the ability to do the work, but the ability to effectively describe how the work should be done to an AI.

Skydive’s business model is designed to encourage this new dynamic. Pricing is based on the compute and AI models the agents use, with no per-user seat licenses and no markup on model costs. This usage-based approach removes the financial barrier to deploying a large number of agents, allowing companies to scale their digital workforce flexibly based on need.

The Governance Gauntlet: Security and Oversight in an Agentic World

Granting autonomous agents their own cloud computers and access to core business systems is an act of immense trust, and one that brings significant security and governance challenges. For Skydive to succeed in the enterprise, it must prove it can provide not just power, but also control. The platform's architecture, which includes centralized administration for access, data, usage, and spending, is a clear nod to these enterprise requirements.

Enterprise IT and security leaders will demand robust guardrails. Key features like sandboxing—isolating agents to prevent unintended system-wide impact—and granular permissions are foundational. Beyond that, a truly enterprise-ready platform requires comprehensive audit trails to trace an agent's actions, strict enforcement of least-privilege data access, and robust human-in-the-loop oversight for critical decisions. These mechanisms are essential for establishing accountability and ensuring compliance with data protection regulations like GDPR and CCPA.

The ethical dimension is equally important. As agents make more autonomous decisions, questions of bias, fairness, and transparency become paramount. Building trust in these AI coworkers will require a clear understanding of why they make the decisions they do. While Skydive is at the forefront of democratizing agent creation, it, like all platforms in this space, will be measured by its ability to navigate this complex governance gauntlet.

With its public launch, Skydive is issuing a bold challenge to the established order of work. It envisions a future where every employee is an AI-empowered force multiplier, capable of building a team to conquer the tasks that previously fell by the wayside. As Dhruv Amin predicts, "By this time next year, I predict most of the job market will have the experience of building their own agents to take on work."