AI agents, or generative AI models, can autonomously perform tasks like transferring information from an email to a spreadsheet. While these models are often touted as productivity enhancers, their propensity for errors suggests they are still evolving. However, many founders, analysts, and investors are optimistic about the potential of AI agents in generative AI. Among these innovators are Bella Liu and William Lu, who founded Orby AI, a platform designed to automate various business workflows.
Orby AI, founded by Bella Liu and William Lu, focuses on automating business workflows such as data entry, document processing, and form validation. While several startups and established companies also offer automation tools, Orby AI differentiates itself with its real-time learning capabilities and its understanding of patterns within unstructured enterprise data.
Liu explains that Orby’s platform observes how employees perform their tasks to create automations for complex activities requiring reasoning and understanding. An AI agent on a worker’s computer watches, learns, and generates automations, continuously adapting as it gathers more information.
Liu, previously at IBM and UiPath, led AI and automation efforts, including product planning and AI-related mergers and acquisitions. Lu, a former systems engineer at Nvidia and engineering lead at Google Cloud, contributed to the development of generative AI for document and database extraction. Orby AI, emerging from stealth mode in 2023, aims to abstract low-level decisions made by workers, allowing them to concentrate on more significant tasks.
The Technology Behind Orby AI
Orby AI fine-tunes its cloud-based generative AI model for specific customer tasks, such as validating expense reports. It incorporates symbolic AI, which uses rules like mathematical theorems to solve problems. While symbolic AI can be inflexible and slow with large datasets, recent research indicates it can be scalable when combined with traditional AI architectures.
Liu highlights that Orby’s model adapts to workflow changes, such as app UI updates, by analyzing API interactions and browser usage. To address privacy concerns, Liu assures that Orby does not store customer data; it only uses the data to fine-tune its model, ensuring encryption in transit and at rest. Additionally, humans are kept in the feedback loop to maintain oversight.
Market Position and Future Prospects
Orby recently secured $30 million in a Series A funding round led by New Enterprise Associates and Wing, with a post-money valuation exceeding $100 million. Despite a competitive market with giants like OpenAI, Anthropic, Amazon, and Google, Liu believes Orby can succeed with a systematic market approach. The company, already generating revenue from about a dozen customers, plans to use its $35 million funding to expand its team and enhance its go-to-market, customer support, product, and technical organizations.
Orby AI’s mission is to meet the enterprise market’s growing demand for generative AI solutions that significantly enhance business performance. By automating complex workflows and adapting to changes in real time, Orby aims to position itself as a leader in the enterprise AI space, demonstrating the practical benefits of AI in improving business processes.
See also: Illumex Eases Data Preparation For LLMs With Generative AI
