Box

Today we reported our third quarter fiscal 2026 financial results. We are proud to have delivered another quarter of strong execution as we exceeded guidance for several metrics and delivered quarterly revenue growth of 9% year-over-year. Customer demand for enterprise-grade AI capabilities are translating into strong momentum for Box AI and Enterprise Advanced. Box CEO Aaron Levie said: "Our strong third quarter performance proves how quickly enterprises are embracing Box as their AI-powered Intelligent Content Management platform. The power of Box is that customers can have a single agentic platform for managing and working with their unstructured data, leverage all of the leading AI models, and have integration across their tech stack." See the full earnings announcement here: https://lnkd.in/gCs4sx2k

Box is the foundational infrastructure for Agentic AI in the enterprise. Through our expanded partnership with Amazon Web Services (AWS), we're enabling AI agents to securely access, understand, and act on your organization's content. Our Box MCP Server, now available through AWS Marketplace, seamlessly integrates with Amazon Quick Suite, Amazon Q Developer, Amazon Strands, Kiro and Amazon Bedrock AgentCore. With AWS' Nova models powering our AI platform, you can unlock AI-driven workflows, search, summarization, and automation directly on your content. We’re at #AWSreinvent this week, where we'll share more about how this collaboration makes enterprise AI accessible, secure, and practical, especially for companies managing regulated or sensitive data. Curious what your first AI-powered workflow could look like? Read more on our blog → https://lnkd.in/gMnWiWPu

Box AI + GPT-5.1 can now analyze financial spreadsheets and scanned PDFs together — and deliver reconciled, auditable insights in seconds. In this demo, Box AI: 📊 Compares “Treasuries” totals across Excel + PDF financial reports 📈 Calculates blended returns from a Performance Analysis PDF 🏛️ Identifies the state with the lowest income tax rate ⚖️ Reconciles discrepancies with transparent citations This is what next-generation enterprise reasoning looks like. Watch the full demo and subscribe to the AI in Action series: https://lnkd.in/g7EfUxJW

We just finished testing Anthropic's Claude Opus 4.5 in Box AI, and the results are striking. When we ran it through complex reasoning tasks and real enterprise workflows, high-effort reasoning hit 83% accuracy, up from 63% in the previous generation. The improvements showed up across industries: 🔹 Education teams jumped from 76% to 96% accuracy on grant compliance analysis 🔹 Energy teams went from 75% to 89% on safety-critical regulatory reviews 🔹 Healthcare teams improved from 42% to 66% when comparing clinical guidelines What makes this different: Claude Opus 4.5 introduces an effort parameter that lets you dial reasoning depth up or down based on what the task requires. High effort unlocks deeper reasoning and self-correction for high-stakes decisions. Medium effort still delivers 79% accuracy but responds faster for everyday work. Instead of one-size-fits-all AI, you get a model that matches its computational intensity to the complexity of the problem. Claude Opus 4.5 is available now in Box AI Studio and through our APIs. Check out our full evaluation here: https://lnkd.in/gPbG8ECn

If your team is still reviewing every grant report by hand, you’re leaving impact on the table. Federal leaders are sitting on a goldmine of insights: four decades of unstructured country plans, programs running in more than 80 countries, and thousands of grant reports that rarely get analyzed end-to-end. In our Box Public Sector Summit fireside, former Federal CIO Jason Gray, shares how AI can: 🔹 Rapidly review grant submissions to see who’s actually delivering on outcomes 🔹 Spot patterns across similar grants so you can scale solutions that work 🔹 Mine decades of reports to replicate successful programs across regions and sectors This isn’t about shiny tech. It’s about using AI to get to “what’s working, where, and why” in seconds instead of months. Watch the on‑demand session to see how AI‑powered grant analysis can change how your agency scales impact: https://lnkd.in/gsc6ChST

We're excited that our new Box AI Agent is now available in Gemini Enterprise and the Google Cloud Marketplace! Watch how it works below👇

Teams don’t need “smarter chatbots” — they need collaborators that actually do work alongside them. In our Partner Podcast series, Box CTO Ben Kus and LangChain CEO Harrison Chase dove into how that’s finally becoming real: 👉 Early “agents” were just 1–2 LLM calls. Now deep agents can run 30+ tool calls to produce a first‑draft PR or research report. 👉 Writing 5–10 line prompts isn't enough; the agents that work best run on 20–30 page instruction sets plus rich tool descriptions. 👉 The real shift is UX: sometimes you want a 1–2 second answer; other times you’re happy to wait while an agent does real work in the background. Manage AI agents like teammates, not tools. Watch the full conversation: https://lnkd.in/gaYDY3dq

Gemini 3 Pro isn’t just another LLM — it’s a step-change in enterprise reasoning. According to Box’s own evals, this model advances multi-step logical thinking, letting you extract and combine information from complex documents with unmatched nuance. In this demo, we drop Box AI (powered by Gemini 3 Pro) into a Hub filled with autonomous driving research and ask: Given the tradeoffs across papers, how should I build a 3-sensor 3D detection stack for an embedded platform? The result? A deeply reasoned, benchmark-conscious architecture recommendation — synthesized in seconds. That’s enterprise automation reimagined: smarter, faster, more context-aware reasoning — powered by frontier AI inside Box. Watch the full demo and subscribe to the AI in Action series for more: https://lnkd.in/gXYJGA-i

70% of bank tech budgets go to maintaining legacy systems while clients demand instant access and seamless collaboration. Here's the shift: 33% of investment banks already have AI in production, with the VDR market set to grow from $2.9B to $7.6B by 2033. Smart banks are using AI to extract insights from unstructured data—turning pitch decks and meeting transcripts into strategic advantages. Discover how AI-powered content management is transforming dealmaking: https://lnkd.in/g3pXa_Fk