Core · Hybrid - Europe
Coders had their Cursor moment, analysts deep research, marketers GPT-4o image generation, and sales teams AI SDRs. What about the rest of the enterprise workflows?
As companies grow, tedious and repetitive operational tasks grow. In fact, most workflows eventually become like this. That’s why businesses outsource them to BPOs—a $300 billion market growing 10% every year—where the work is offshored to low-income labor. Alternatively, they try to automate it with legacy, fragile tech like UiPath (RPA). Because of the low performance of these solutions, they end up automating only a fraction of these workflows.
That was before computer-use agents—agentic systems that can automate any workflow in any application by controlling a browser or a computer.
At Twin we pioneered computer-use agents in early 2024. Since then we’ve shipped thefirst fully autonomous agent—Invoice Operator—and rolled it out to 500k SMBs via our first enterprise customer, Qonto. And it actually works: we achieved 96% accuracy on an internal benchmark of 10 000 invoice-collection tasks.
We now have strong traction in this first fintech vertical and are already expanding into others with our platform, the first that enables enterprises automate processes with fully autonomous agents.
We’re pursuing the $1 trillion agentic process-automation opportunity. Not only do we aim to build a massively successful business, we also want to define how an enterprise SaaS company should operate in the era of agents: small, technical teams augmented by agents.
Twin is the platform for creating, testing and deploying agents, and its core engine sits in the center of the user journey -- users communicate with it to create browsers and tabs, and interact with them remotely through their browser, launch agents, and inspect the results of these agents
A snapshot of the topics of the Core team: