Orby AI Raises $30M in Series A for GenAI Platform for Process Automation

Automation has been a top initiative for IT leaders for decades. However, attempts to automate the tedious, repetitive work that occurs daily in most enterprises have not moved the needle very far. Mountain View, Calif., startup Orby AI says it can make automation a reality using generative AI.

Bella Liu, co-founder and CEO of Orby AI, says the company is tackling a significant issue that has long hindered enterprise productivity. She points out that traditional automation solutions, like most software, are rule-based. This means that if a team wants to automate a process, they must create rules of the “if this, then that” variety, a time-consuming process, before they get to the actual work they want to accomplish.

So, the Orby team decided to do away with the rules and take a unique approach to the problem.

No Rules

“Our approach uses what we call “observe, learn, automate, and maintain,” said Liu, who was director of product management – AI for UiPath before she co-founded Orby with CTO Will Lu in 2022. “We set out to use GenAI to build a platform that could change how enterprises handle repetitive tasks. Repetitive work is an efficiency killer for organizations. It's error-prone and negatively impacts employee job satisfaction.”

Liu said process automation has been built on explicitly defined rules—until now. Orby is approaching the issue from the opposite perspective, which the company calls “no rules.” “The rules used in today’s robotic process automation (RPA) are very limited and fragile,” said Liu. “If you have a huge implementation project, automating one process takes a lot of time. And you have to rely heavily on IT developers to build it out. That’s why it’s so hard to scale automation to all the processes you want to automate.”

Orby built the first generative process automation (GPA) platform to tackle automation's many challenges by leveraging artificial intelligence's capabilities. “Users install Orby as an AI agent on their computers and show it one iteration of a repetitive process,” explained Liu. “Orby will observe and show, in real time, how well it understands the task. Orby will create a workflow automatically and ask if it’s right. Then, the human user needs to click a button to confirm. And now, that process is automated.”

Liu said Orby applies continuous learning to improve its work. If Orby’s confidence level is low for certain steps as it runs the automation for the user, it will ask for feedback, such as ‘Is this right?’ or ‘Should I do it differently?’ The user clicks a response button. Orby gathers the feedback and takes over the rest of the process. Orby then becomes smarter and better suited for each user’s processes.”

The Critical Role of Large Action Models

Orby utilizes patented GenAI technology with intelligent AI agents and a first-of-its-kind multimodal large action model (LAM).

“Our large action model is informed by what it sees the user doing, where they're clicking, what other applications they’re accessing, what the HTML is on the page, where they're clicking on, etc.,” said Will Lu, co-founder and CTO of Orly. “It learns and creates application automation without IT having to be involved in coding and understanding the business process.

Lu co-founded Orby after leaving Google, where he was head of engineering and focused on cloud AI. Earlier, he was a software engineer with NVIDIA. “Because often there's a disconnect between the business users, the people doing the work, and the IT group that has to understand, ‘Okay, what are you doing? What's your workflow? How do I code something.’ A lot of the workers don't know what's technically feasible, number one, or number two, they don't necessarily understand how to describe their workflow to a point where they can get IT involved to create an automation,” he said.

The Orby founders say their GPA solution is gaining traction in finance and accounting departments, including complex auditing use cases. “A Fortune 500-type customer receives millions of receipts yearly that employees want to expense,” said CEO Bella Liu. “The auditing team needs to look through all the expenses and ensure all the data matches the receipts and that the expenses comply with all the internal rules. It's very time-consuming, at about 15 minutes per receipt. And so, they don't check all of them. They only do simple challenges because it's very complex and too time-consuming.

“With Orby, they can automate 100% of the auditing. The auditors need to show the platform what they do. Then, Orby generates a workflow. The auditors can begin to run it and give feedback to Orby so it learns.”

Leading VCs Believe in Orby

From an initial seed investment and the Series A round announced this week, Orby has raised $35 million. Investors include Pear Venture Capital, New Enterprise Associates (NEA), Wing VC, and WndrCo. This follows the 97.2M round of funding that Emergence AI received earlier this week. AI has been viewed as a massive change agent for employee and customer experience.

VC investment has undoubtedly aligned with this thesis, and I expect to see more announcements like this throughout the year.

Zeus Kerravala is the founder and principal analyst with ZK Research.

Read his other Network Computing articles here.

Related articles: