Box Inc., the content management company that helps organizations to manage their files in the cloud, says it’s expanding its artificial intelligence capabilities to more users and adding a new mechanism for tracking AI-related costs, so customers can keep tabs on their spending.

The announcement came as the company released yet more AI features into its product, including Box AI Extract agents and Multi-Doc Querying capabilities, to help users unlock more insights from various types of documents.

In a blog post, Box Vice President of Product Management Yashodha Bhavnani said the recent breakthroughs by companies such as DeepSeek Ltd. are ushering in a new era of cheaper AI. Now, the company wants to capitalize on that by helping its customers optimize their AI spending, so they can maximize their investments in the technology.

Box has strived to make AI a central feature of its content management platform, integrating the technology into every aspect of its products, from the core metadata layer to the user experience. In 2023, it launched a suite of generative AI tools called Box AI for Enterprise Advanced and Enterprise Plus subscribers, saying it wanted to help companies unlock the value hidden within their content and make every worker be more productive.

Since then, Box has progressively built on its AI capabilities. Last year it announced the launch of Box AI Studio, a platform for customers to create and use AI agents that can handle more complex tasks for users, with use cases in employee onboarding, customer service and more. It also announced Box Apps, a no-code development studio for creating intelligent applications for business processes such as contract management and invoice processing.

Having firmly established its AI capabilities, Box is now making them available to more customers. As of today, subscribers on Box’s Business, Business Plus and Enterprise plans can access the Box AI features for the first time. Moreover, it’s adding a new feature called Box AI Units, a new mechanism for tracking AI usage across organizations.

Bhavnani said Box Units will prove extremely useful as customers increase their reliance on its AI features. He explained that traditional pricing models often lack transparency, so users fail to understand how much they’re spending on AI, and cannot easily measure the value they receive from it.

Box Units aims to provide more clarity for customers on their AI spending. It’s a new mechanism for tracking and managing AI usage that quantifies each interaction, whether it’s querying content or extracting metadata or something else, into discrete units. In doing this, businesses can more easily understand how they’re using Box’s AI functionality, Bhavnani said. In turn, this will make it easier for them to forecast their AI costs and optimize their budget accordingly, he said.

Expanded AI capabilities

The new price tracking tool will be all the more useful as Box continues to expand its AI offerings. The new Box AI extract agents announced today are meant to accelerate decision-making and automate critical workflows, and they also provide users with the ability to create custom extract AI agents using the Box AI Studio application programming interface.

The extract agents are designed to pull key details from any type of document, including images, without someone needing to manually tag those files first. Instead of tagging, the AI agents learn by themselves to understand what’s in those files, so they don’t need any help. As an example, Bhavnani said a property management company could use an extract agent to pull key information, such as contract expiration dates and details on rents from their customer agreements, enabling faster renewals, analysis and review processes. They’re able to extract key info, validate the accuracy of that info, and then save key details into the metadata to make it more readily available for other users.

Customers will also be able to leverage Box AI extract agents through the company’s API, giving them more control. Users will be able to define customer fields and instructions for each extract AI agent, fine-tuning them for specific scenarios.

In addition to the agents, the company announced a new image querying feature (below) that makes it possible for users to ask questions about image files stored in Box, including photographs, graphics, document scans and blueprints.

According to Bhavnani, this can unlock huge productivity gains for users, giving them access to instant insights from thousands of images. As an example, a property rental company could use it to quickly assess and understand the turn-in condition of one of its properties, while insurance firms might use it to assess the damage to a car that was involved in a collision.

Finally, the company announced a similar tool called Multi-Doc Querying, where users can select multiple files at once and ask questions about them. The feature is somewhat similar to the Box AI for Hubs feature that was unveiled last year, enabling users to ask questions of their document hubs, which are sets of related files in a specific folder.

Bhavnani said the tool eliminates the need for employees to spend time manually checking images, so they can find the information they need much more quickly.

Featured image: SiliconANGLE/Dreamina