AI “agents” are generative AI models that can perform actions autonomously, like copying info from an email and pasting it into a spreadsheet, and have been hailed as productivity superchargers. That might be a bit premature, given models’ tendency to make mistakes. But at least a few founders (and analysts and investors) seem convinced that agents are the next frontier in generative AI.
Bella Liu and William Lu are two such founders. Their company, Orby AI, is building a generative AI platform that attempts to automate a range of different business workflows, including workflows that involve data entry, documents processing and forms validation.
Lots of startups offer tools to automate repetitive, monotonous back-office business processes (see Parabola, Tines, Sam Altman-backed Induced AI and Tektonic AI, to name a few). Incumbents, too, like Automation Anywhere and UiPath, have moved to embrace AI to try to maintain pace with the generative AI competition.
But Liu and Lu claim that Orby’s tech stands out for its ability to learn and act on workflows in real time and to understand the patterns and relationships within an enterprise’s unstructured data.
“Orby’s platform observes how workers do their work in order to automatically create automations for complex tasks that require some level of reasoning and understanding,” Liu, Orby’s CEO, explained. “An AI agent installed on a worker’s computer effectively watches, learns and generates automations, adapting the model as it learns more.”
With Orby, which launched out of stealth in 2023, Liu and Lu say that they sought to create AI that could understand some of the low-level decisions being made by workers and abstract those decisions away, freeing up workers to focus on headier things.
Liu previously led AI and automation efforts at IBM, including product planning and AI-related mergers and acquisitions, and was UiPath’s director of AI product management. Lu is a former Nvidia systems engineer who joined Google Cloud as an engineering lead, helping to design generative AI document and database extraction tech.
Orby’s purported secret sauce is a cloud-based generative AI model that’s fine-tuned to complete customer tasks, such as validating expense reports. The model relies partly on symbolic AI, a form of AI that leverages rules, such as mathematical theorems, to infer solutions to problems.
Symbolic AI alone can be inflexible and slow, especially when dealing with large and complicated data sets. It needs clearly defined knowledge and context to perform well. But recent research has shown that it can be scalable when paired with traditional AI model architectures.
“For the last two years, we’ve been engineering this AI model, and have performed successful trials,” Liu said. “There are few pure-play generative AI companies attacking the enterprise head-on with something end-to-end. We are one.”
Liu says that Orby’s model can intelligently adapt to changes in workflows, like when an app’s UI gets an update, by analyzing API interactions and a worker’s browser usage. Having software monitor an employee’s every move sound like a privacy disaster waiting to happen. But Liu claims that Orby doesn’t actually store customer data; it only uses it to fine-tune its model, encrypting the data both in transit and at rest.
“Humans are kept completely in the feedback loop,” she added.
Orby, which recently raised $30 million in a Series A funding round co-led by New Enterprise Associates and Wing, sources say at a post-money valuation north of $100 million, is competing in a challenging sector. Forthcoming agentic AI from generative AI powerhouses such as OpenAI and Anthropic have dampened the prospects of incumbents and smaller players alike.
Adept, a startup building AI agents technology focused on enterprise applications, is reportedly on the cusp of an acquihire deal with Microsoft before it manages to ship a single product. Amazon and Google have released AI agent tooling to little fanfare. Elsewhere, UiPath — despite its ramping up of generative AI initiatives in the past year — saw sales plummet in its most recent fiscal quarter.
Liu says that Orby can come out ahead by taking a systematic go-to-market approach. The company is already generating revenue from around a dozen customers, she says, and plans to put its $35 million war chest toward expanding its Mountain View-based, roughly 30-person team.
“The funds are being used to scale our go-to-market, customer support, product and technical orgs,” she said. “The enterprise market has an insatiable appetite for generative AI solutions that demonstrably improve business performance; they are just trying to figure out where to best apply the technology in the near term before they scale it across their business.”