Box ushers in a new era of AI integration with Box Automate

On Thursday, Box opened its developer conference Boxworks by announcing a new range of AI features and integrating agent-based AI models into the company's products. These announcements reflect the increasing pace of AI development at Box. Following the introduction of AI Studio last year and a series of new data extraction agents in February, and followed by additional AI agents for search and in-depth research in May, Box now presents a system called Box Automate. This system acts as a type of operating system for AI agents and enables the segmentation of workflows, which can be supplemented by AI as needed. In a conversation, CEO Aaron Levie explained the company's strategic direction regarding AI. While emphasizing the opportunities of AI agents in the modern workplace, he also provided clear insights into the current limitations of the models and how these can be overcome with existing technology. Levie elaborated that automating workflows involving unstructured data is crucial, as previous automations were limited to structured data. Through AI agents, unstructured data can be accessed effectively for the first time. At the introduction of Box Automate, Levie emphasized that customers need to be convinced that workflows are executed with the desired accuracy and predictability. Segmenting processes into different agents provides the necessary control and flexibility in this regard. Additionally, Levie discussed the importance of security and data control. With a powerful layer for access control and data security, Box aims to ensure that only authorized individuals can access necessary information. In the face of competition from base model companies, Levie stated that Box distinguishes itself through comprehensive security and integration options with leading AI models on the market. The ability to integrate and manage a variety of AI models without being tied to a single platform is a key differentiating feature of Box.