← All organizations

Enterprise content management and AI workflow automation

Box

Box provides enterprise content management for teams that need to store, share, govern, and work with business files. Box Agent searches across organizational content, analyzes documents, and creates deliverables from natural-language requests. Box AI Studio lets administrators configure custom agents and choose their underlying models, while Box Extract turns unstructured files into structured metadata. Box Forms, Doc Gen, Sign, and Automate connect information collection, document creation, electronic signatures, and automated workflows. Developers can extend the platform through APIs and SDKs.

Co-founders Aaron Levie and Dylan Smith have served as CEO and CFO, respectively, since 2005. Box has expanded from file sharing and collaboration into AI-assisted content workflows. Its technical approach keeps content permissions and governance within the platform as people, applications, and agents access files. Its MCP server connects external AI tools to Box content, supporting multiple model ecosystems rather than a single underlying model.

Box’s engineering publications illustrate its operating scale: in 2026, the company reported peak traffic exceeding 4,000 file uploads per second. For folder-size aggregation, its engineers combined bounded session windows with inactivity gaps in Dataflow. Box reported that this reduced peak database writes from 24,000 per second to 6,000–8,000 and saved approximately $300,000 annually in database infrastructure costs.

www.box.com

1 talk

Newest first

1 speaker at AIE

Affiliations reflect their AIE appearances, not necessarily current employment.

Messages from the stage

Prompts and workflows over fine-tuning

Kus describes a preference for prompts, prompt caching, and agentic workflows over fine-tuning when developing document extraction.

Affiliations reflect each recorded session, not necessarily current employment.

Company sources · checked 2026-08-28