In 2025, the focus is on AI agents that automate tasks using generative AI technology. A new player in this space, Orby, introduces its Large Action Model (LAM) designed specifically for enterprises. Unlike traditional AI models, LAMs input actions from various sources like application screenshots and user interactions to streamline workflows. Orby’s co-founder, Will Lu, notes that their LAM identifies and automates complex tasks within software like Salesforce. By collecting workflow data, or “traces,” Orby’s model improves automation efficiency. For CIOs, Lu emphasizes the importance of addressing real business challenges, ultimately helping organizations save time and enhance productivity without the need for direct API integrations.
In 2025, we are witnessing the rise of AI agents, which utilize generative AI to automate various tasks. A new player in the enterprise arena, Orby, is introducing a concept called Large Action Model (LAM). Unlike traditional large language models (LLMs) that focus mainly on generating text or images, LAMs are specifically designed to automate processes within the business world. Will Lu, co-founder and CTO of Orby, explained that LAMs take actions like application screenshots and user interactions as inputs, enabling them to handle complex workflows effectively.
Orby’s LAM, named ActIO, has collected over a million action sequences, allowing it to understand and streamline multiple enterprise tasks. The model is trained both on open web data and customer-specific data, providing a powerful foundation for automating business processes.
One notable aspect of Orby’s approach is its grounding concept, which focuses on pinpointing UI elements necessary to execute actions within software applications. This feature was developed through a collaboration with Ohio State University and enhances the ability of AI to navigate complex systems without requiring API integrations.
With AI agents becoming essential tools for enterprises, Lu advises CIOs to focus on identifying real business pain points. Tasks that consume significant time for employees can often be handled easily by AI, such as auditing expense reports—a process that traditionally involves tedious checks against company policies.
Security remains a cornerstone of Orby’s offering, ensuring that enterprises can adapt AI solutions without compromising their systems. Even as Orby seeks to automate extensive workflows, it maintains a human oversight process, acknowledging that AI technology is not yet foolproof.
In summary, Orby’s innovative use of LAM technology demonstrates the transformative potential of AI in improving efficiency and productivity for enterprises.
Tags: AI agents, Large Action Model, enterprise automation, Orby, generative AI, workflow optimization, business technology, security in AI.
What are AI agents in the enterprise?
AI agents are computer programs that can perform tasks for businesses. They use artificial intelligence to help with things like improving customer service, analyzing data, and automating routine tasks.
How do AI agents help businesses?
AI agents save time and reduce costs. They can handle repetitive tasks, analyze trends quickly, and provide insights that help decision-making. This lets employees focus on more important work.
What tasks can AI agents automate?
AI agents can automate various tasks, such as data entry, scheduling meetings, customer inquiries, and even financial reporting. This automation increases efficiency and accuracy in daily operations.
Are AI agents safe to use in the workplace?
Yes, when properly designed and monitored, AI agents can be safe. Businesses should keep data privacy and security in mind and ensure that AI systems follow the company’s guidelines and regulations.
Will AI agents replace human jobs?
AI agents may change the way some jobs are done, but they are more about assisting humans than replacing them. They can take over simple tasks, allowing workers to spend their time on more complex and creative jobs.
