News
In Brief
- Box Extract AI-powered tool extracts data from unstructured documents automatically.
- Converts document content into searchable, exportable business metadata.
- Integrates extracted data with enterprise systems like Salesforce, Snowflake.
Box, Inc. announced the general availability of Box Extract on Jan. 15. Using AI models from Google, Anthropic, and OpenAI, the tool finds data buried in unstructured documents and turns it into metadata.
Box Extract incorporates agentic capabilities that enable the tool to understand document structure and meaning, breaking content into components such as paragraphs, tables or charts, according to the company. Custom Extract Agents are available through the Enterprise Advanced plan in two tiers: Standard Extract Agent for simple data capture and Enhanced Extract Agent for complex, multimodal documents.
The extracted metadata integrates with Box Apps, Box Relay and future Box Automate workflows, and can sync to external systems including Databricks and Snowflake. The company is positioning the solution specifically for heavily regulated industries dealing with high-volume document processing where accuracy, security and compliance are critical.
Box Extract Capabilities
| Capability | Description |
|---|---|
| Box Extract | Extracts data from unstructured content using generative AI models |
| Standard Extract Agent | Streamlines simple data capture for faster, cost-efficient results |
| Enhanced Extract Agent | Handles complex, multimodal documents with deeper reasoning |
| Custom Extract Agents | Tailored agents deployable at scale across content types |
| Metadata Integration | Stores structured data alongside content; syncs to Databricks, Snowflake, Salesforce |
Box Builds Its AI-Native Content Intelligence
Box has maintained an aggressive pace of AI integration throughout 2025. The company started 2025 with the unveiling of its Enterprise Advanced Plan earlier in 2025, introducing Box AI Studio for building custom AI agents. It rapidly incorporated new large language models as vendors released them throughout the year.
In mid-May 2025, the company unveiled an all-new AI platform featuring AI Agents for search, deep research and data extraction. Strategic partnerships expanded Box's reach into regulated markets: in March 2025, Box achieved FedRAMP High authorization for U.S. federal agencies and announced a collaboration with IBM to run Box AI on IBM watsonx models in April 2025.
Financial results reflected growing momentum: Q3 2026 revenue reached $301 million, a 9% year-over-year gain.
AI-Driven Document Intelligence Reshapes Enterprise Workflows
Generative AI is changing how enterprises extract intelligence from unstructured data, automating workflows while raising new security and compliance concerns.
The technology has moved intelligent document processing beyond simple data extraction into document understanding. Enterprise platforms have pivoted to leverage unstructured data, as email messages, chat transcripts and meeting notes now represent the majority of organizational information.
AI-driven automation can generate targeted workflows ready to test in minutes versus traditional manual configuration. Recent data indicates that 88% of enterprises plan AI agent initiatives within six months. These developments are reshaping the digital workplace as organizations seek to unlock value from previously inaccessible content.
Have a tip to share with our editorial team? Drop us a line:
