Orby AI is pioneering what it calls generative process automation, an approach to enterprise workflow automation built around a Large Action Model (LAM) rather than the brittle, rules-based scripts of legacy robotic process automation. Where traditional RPA requires engineers to painstakingly specify each step, Orby's agents simply observe users at work, learn which tasks are repetitive and automatable, and then generate the code and actions needed to execute them.
The core insight is that the highest-value automation in the enterprise is hidden inside the day-to-day actions of knowledge workers, the copy-paste, data-entry, reconciliation, and multi-system workflows that consume hours but were never documented. Orby's LAM is trained to understand these actions across applications, identifying patterns and proposing automations without a human having to map the process first.
This observe-learn-automate loop positions Orby distinctly against both classic RPA vendors and pure LLM chat tools. By grounding its model in real user actions and producing executable automations, Orby aims to deliver measurable workflow replacement, exactly the kind of outcome enterprise buyers increasingly demand from agentic AI rather than open-ended autonomy claims.
Orby was incubated with backing from leading firms and raised a $30 million Series A co-led by New Enterprise Associates (NEA), Wing Venture Capital, and WndrCo, with participation from Pear VC. The funding is aimed at accelerating development and commercialization of its generative process automation platform.
For enterprises drowning in tedious, repetitive back-office work, Orby offers a path to automation that scales with how people actually work, continuously discovering new opportunities rather than waiting for IT to script them, and turning the Large Action Model into a practical engine for enterprise efficiency.