The enterprise software model is fundamentally broken. Software decays, SaaS is bloated, consultants create dependency. The next generation of enterprise software will be autonomous, continuously learning, and adaptive, but every enterprise system today is built for humans doing the work. AI agents can now reason, act, and learn, and every enterprise I talk to is racing to deploy them. The problem is that the platforms beneath those agents were never designed to support the enterprise.
After twenty years of building and scaling platforms at Sun, VMware, Groupon, and Atlassian, I’ve seen these issues up close. On the investing side at Coatue, I saw the size of the agent readiness problem over and over again.
That’s why I started Sycamore: to build the operating system for the AI-native enterprise, where software finally builds itself.
Patterns: What I Keep Seeing
Three patterns keep compounding across the enterprise.
Software decays. Teams end up spending more time maintaining what they shipped last year than building what they need next year. The moment software is deployed, it becomes brittle and stale, bugs linger, and integrations drift out of sync with how people actually work. I’ve seen firsthand how even the most sophisticated engineering organizations struggled to keep their software ecosystems healthy. We’ve all accepted this decay as the cost of doing business, but I believe it doesn’t have to be that way.
SaaS has become bloated. The promise of SaaS was simplicity, but the reality is that most enterprise products ship with fifty features when the customer only needs five. Organizations end up adapting to the software instead of the other way around. We’ve traded one form of rigidity for another.
The consulting model creates dependency. Many large enterprises I’ve worked with have spent millions on consultants who shoehorn solutions into existing systems and then move on, placing the burden of managing that custom-built work back on IT organizations that are already short-staffed and overwhelmed. Instead of building capability within the organization, this model creates ongoing dependency.
AI agents that can reason and act are the next major platform shift in enterprise computing. At Coatue, I saw enterprise after enterprise spinning up agent proofs-of-concept, but they all hit the same walls. There was no infrastructure to govern these agents, no system to operationalize them, and no secure way to move from demo to production at scale.
At Atlassian, we grew the engineering team from 800 to nearly 4,000 people. We transitioned the entire company to a cloud-first platform, and through all of it, I learned something that has stayed with me: before you can scale any product, you have to build the platform beneath it. That cloud foundation took time to build, and it only succeeded because we made it the priority, not an afterthought.
The enterprise AI moment requires the same thing. Sycamore is building that foundation.
Software That Builds Itself
Sycamore is where software builds itself. Our core vision is to redefine what enterprise software looks like in the AI era. Our mission is to help every single organization radically transform and bring AI into the company to maintain their competitive edge, ensuring they become stepwise more effective.
Instead of getting stale, Sycamore software starts improving immediately. It fixes its own bugs, personalizes itself to the user and to the organization, and runs around the clock, continuously getting better, all without the massive overhead of traditional custom work.
Three things make this possible.
The Agent Operating System is the kernel beneath everything, the foundation that manages the entire system and drives continual improvement. It watches every decision an agent makes, learns what’s working and what isn’t, continuously adjusts, and rewrites code and software on the fly. We’ve brought rigorous CI/CD discipline to the product and agents themselves, ensuring that code and features are tested robustly before being deployed.
At Atlassian, I learned that at scale, you can’t move fast without trust. We spent years building that foundation for cloud infrastructure, and it’s even more critical for autonomous agents. So we built trust into Sycamore from day one, rather than treating it as something to add later. Every operation is isolated, auditable, and governed.
The Adaptive Factory is the engine that generates software tailored to each organization and its needs. It doesn’t just automate workflows, it rethinks them from first principles. It steps back and asks a different question: what’s the real problem that needs to be solved? Then it designs and builds the streamlined solution from scratch, whether it requires AI agents, backend software, frontend software, or data connectors.
Collective Intelligence is what connects everything to the organization itself. Sycamore learns your company’s values, hierarchy, and institutional knowledge. It understands that a marketing manager and a CEO should never get the same answer to the same question.
All of this is built on an enterprise-first foundation. Sycamore is designed from the ground up for the requirements large organizations care about most: trust, security, compliance, and auditing. Our agents respect roles, permissions, and control planes.
How We Operate
In order to deliver the autonomous enterprise, we have to live it ourselves.
Sycamore runs at machine speed. Our agents work around the clock, and we’ve moved from two-week sprints to daily sprints. In the morning, our team plans and collaborates with our agents. By evening, the agents are set up to do the work while the team is sleeping. We ultimately measure our agents’ productivity, not our engineers’.
It’s not just the engineering team. We encourage every single person in the organization to build using Sycamore. We are committed to simplicity and the fact that voice is our default interaction mode is testament to this.
This goes beyond how we work together as a team. When we kick off with customers, our agents have already done the homework and this ensures a personalized and more informed experience from the very first engagement.
That’s the kind of incredible experience we believe every enterprise should be able to offer and it’s what Sycamore enables.
Who’s Building With Us
We’re already working with Fortune 500 enterprises as design partners to deploy autonomous AI agents in production. The feedback from these early deployments is shaping every layer of the platform.
Our team includes researchers from Stanford and Cornell and engineers from Meta, Google, Atlassian, and other leading technology companies. We’ve built enterprise platforms at global scale before, and we’re applying that experience to the AI challenges that every enterprise will face in the next few years.
Today, we’re announcing a $65 million seed round led by Coatue and Lightspeed Venture Partners, with additional participation from Abstract Ventures, Dell Technologies Capital, 8VC, Fellows Fund, and E14 Fund. We’ll use this investment to grow our engineering and applied AI teams, go deeper with our enterprise partners, and push forward on the hardest research problems in trust, memory, and multi-agent coordination.
What’s Next
We believe that the future of work involves making companies autonomous, but with humans in control. Every employee should be able to create, build, and solve problems with AI, whether they write code or not. The software should be smart enough to harness the creativity of the entire organization and trustworthy enough that enterprises can give it real responsibility. Sycamore is building the operating system for that future, with a foundation of trust, security, and control.
If you’re an enterprise leader working through how to move AI agents from pilot to production, we’d love to talk. And if you’re an engineer, researcher, or builder who’s passionate about the problems and wants to help solve them, we’re hiring across the board.
We have an amazing team, an incredible set of partners, and a mission that matters. I couldn’t be more excited about what we’re building.
Learn more at sycamore.so. Follow us @SycamoreLabs on X.