``Right now, the most noteworthy thing is the behavior of AI agents.Currently, they are used to perform limited tasks, but in the future AI agents will automate complex workflows,'' he said (Figure 1).
He further explained, ``AI agents currently perform a variety of tasks such as creating, reviewing, and coding various materials.As AI agents become compatible with agent-type workflows, a wide range of business processes themselves will be transformed.'' (Figure 2)
But, he says, "To make that transformation happen, AI agents need to know everything about your business." He continued, ``This information includes information unique to your company, such as product specifications, research and development, marketing, customer data, and human resources data.These are the very contexts of your business.And they exist in the content you own.Therefore, it is important to link that content with AI agents.''
He went on to explain, ``Now, 90% of business data within companies is unstructured data (such as documents).In the past, computers were designed for structured data (such as financial data), and unstructured data could not be utilized.However, in order to improve the efficiency and further automate operations, it is necessary to effectively utilize unstructured data.'' The company refers to this unstructured data as content.
However, he retorted, saying, ``There are major problems with traditional management methods for unstructured data.'' what is the problem? He explained, ``There are many cases where AI agents are unable to access various content within a company.For some reason, this is because data is scattered and fragmented for each of the various applications used within the company, and cannot be used in conjunction with each other.What companies are looking for is to flexibly and securely link content and AI agents to streamline and automate their operations.'' (Figure 3)
Box's services, which form an intelligent content management platform, make this possible (Figure 4).
The overall picture is shown in Figure 4, which consists of ``infrastructure'', ``data'', ``content services'', and ``AI platform'' from the bottom, and can be accessed flexibly and safely by ``users'', ``AI agents'', and ``applications'' through various connection technologies including UI (user interface). The composition of these three people lined up in the same position on the top row is also unique to Box.
Additionally, Box has long built an extensive partner ecosystem, allowing both its apps and AI agents to work with leading companies from other companies (Figure 5).
Transform business processes with agent-based workflows
At the end of his speech, Mr. Levy showed the process of AI transformation in Figure 6 and said the following:
"There are three steps to the AI transformation process. The first step is to make knowledge work more efficient with AI. The second step is to be able to mine large amounts of information. And the third step is to be able to transform business processes with agent-based workflows. I hope many companies will take these three steps with Box."
What caught my attention in Mr. Levy's explanation of the AI transformation process was the expression ``transforming business processes with agent-based workflows,'' which I mentioned in Figure 2. Isn't this something that goes beyond the framework of conventional content management platforms? The company may be aiming for a position as a platform for agentic AI, which refers to the technology and usage environment that orchestrates and manages what he calls "AI agents" to streamline and automate company-wide operations.
In that sense, Box has an advantage. This is because they have focused on content management platforms and have not worked on business applications, so they are in a neutral position in the agentic AI market that is expected to grow significantly in the future. Box has maintained this position in the SaaS market to date, allowing it to form a broad partner ecosystem. It is my opinion that this unique position will work effectively in the agentic AI market as well.
However, in order for Box to maintain its influence as an agentic AI platform, it will be important to further increase the presence of its original data infrastructure, a content management platform. Just like the agentic AI market, this data-based market is likely to become a hotly contested area in the future.
Under such circumstances, how will Box evolve with AI? I also see this as an epoch-making movement that reflects how the agentic AI market will take shape in the future.
