At BoxWorks, learn from @OpenAI, @nvidia, and @awscloud, alongside a community of developers and technical builders. Join us November 5–6 in San Francisco for technical sessions, interactive demos, and hands-on programming workshops, Limited number of free developer summit passes are still available. Grab yours here: luma.com/box-dev-summit-2026

The Box MCP connector is now generally available in @harvey's Connector Library. Search files, query documents, and create content in Box directly from Harvey, with existing permissions enforced throughout. Watch it in action.👇

Agent swarms could make compute a new lever for tackling complex work. @levie joined @CNBC to explain how groups of agents can work in parallel to unlock productivity in life sciences, finance, software development, and more.

“The systems of record that agents have access to will grow in importance, because you will want to make sure that those systems are governing the rules of the road for these agent swarms to operate in.” @levie on @CNBC discussed why enterprise ROI from agent swarms relies on the systems underneath them.

The countdown to BoxWorks is on, and we can’t wait to bring this community together again. Join the brightest minds in Content + AI November 5–6 in San Francisco for keynotes, hands-on demos, and partner solutions in the AI innovation showcase. Connect with fellow builders and explore how business content can power your next AI workflow. Bring your questions, and leave with ideas to put to work. Register: boxworks.box.com/series/boxw…

Box AI Extract pulls structured fields from documents, and @OpenAI Decisions API routes them to the right workflow. In this demo, a COI arrives in Box through a File Request. Box AI extracts coverage fields from both the upload and the project requirements. The Decisions API receives that evidence and returns a typed route: request endorsement, specialist review, or ready for human review. A Box webhook triggers the whole process on document upload. The app creates a review task with context, adds the right collaborator, and writes extracted values and decision confidence back to the file as metadata. Box is where the documents arrive and where people review them. Watch👇

At the BoxWorks Developer Summit, join developers, AI engineers, and technical leaders for two days of talks and hands-on workshops exploring how to unlock AI with unstructured data. Hear what’s next for the Box platform, build with Box APIs, SDKs, and AI agents, and explore real-world enterprise use cases. Speakers include technical leaders from Box, @nvidia, @OpenAI, @getpostman, and @daytonaio. Registration includes access to both BoxWorks and the Dev Summit: boxworks.box.com/series/boxw…

3 weeks until BoxWorks brings the enterprise AI community together in San Francisco. We’re grateful to our Premier Sponsors @awscloud, @Deloitte, @googlecloud, @IBM, @TCS, @DataBank_Ltd, and @MicroStrategy, and to all our partners helping organizations put Content + AI to work. Join us November 5–6 to meet our partners, explore their solutions in the Expo Hall, and see what’s next for enterprise AI. Register today: boxworks.box.com/series/boxw…

🚀@Microsoft Execution Containers (MXC) provide policy-driven, sandboxed execution across Windows, Linux, and macOS. Box Mount runs inside that sandbox and maps a Box folder to a standard POSIX path, with two-way sync and Box permissions and governance applying automatically. In this demo, the agent reconciles fictional invoices, purchase orders, and payments using deterministic tools. It identifies six exceptions including duplicate records, a missing PO, an over-limit invoice, and unpaid invoices. The finished Markdown review and Excel exception workbook sync back to Box through Box Mount. MXC is the isolated execution layer, and Box Mount is the governed content layer. The result lands in Box, ready for human review. Demo ↓

We put Gemini 4 Argon from @GoogleDeepMind to work untangling an insurance billing dispute across 12 documents and 2 policies. In Box AI, Argon traces a payment credited to the wrong policy and catches a bank return that changes the balance. Argon also reconciles the customer’s expected premium credit with the carrier’s final amount. The statement shows $6,000. The customer expects $4,200. Argon generates a cited review explaining why $6,800 remains due after recommended corrections, with a draft customer response for human review. Build custom agents for your own insurance workflows with Box AI Studio.