gcp Google Cloud Blog ·

Box enhances enterprise agents with Gemini Multimodal Embeddings 2

blogaigcpgaarchitecthealthcarefinance
feature

Box is integrating Google Cloud's Gemini Multimodal Embeddings 2 into its Agentic Platform, significantly enhancing enterprise content management. This enables AI agents to process and understand multimodal content, including visual, spatial, and structured data, alongside traditional text. The integration helps preserve complex document layouts, interpret visual evidence, and bridge heterogeneous file formats like PDFs, spreadsheets, and images. This advancement empowers organizations, particularly in finance and healthcare, to build more sophisticated AI agents for complex workflows, from financial reporting to clinical decision support.

  • Box integrates Gemini Multimodal Embeddings 2
  • Deeper understanding of visual and structured content
  • Improved financial and analytical reporting
  • Enhanced clinical decision support and diagnosis
  • Reconcile data across diverse document formats
Features (1)
  • Box integrates Gemini Multimodal Embeddings 2

    Box's Agentic Platform now incorporates Google Cloud's Gemini Multimodal Embeddings 2, extending its AI capabilities beyond text-based retrieval-augmented generation (RAG). This integration merges Box's industry-leading Intelligent Content Management platform with Google Cloud’s advanced multimodal AI embeddings.

Enhancements (1)
  • Deeper understanding of visual and structured content

    The integration allows AI systems to interpret documents like humans do, preserving the spatial layout of tables, charts, and flowcharts without losing their inherent meaning. It illuminates visual indicators and connects information across diverse file formats such as PDFs, spreadsheets, and images.

Notes (3)
  • Improved financial and analytical reporting

    AI agents can now analyze highly structured financial documents, understanding embedded tables and growth charts while preserving their physical structure. This enables accurate trend analysis and contextual sourcing of specific metrics from complex portfolios for corporate finance, research, and audit teams.

  • Enhanced clinical decision support and diagnosis

    Critical patient data fragmented across visual and textual formats, like clinical photos, pathology slides, and risk matrices, can now be synthesized simultaneously. This aids granular anomaly identification and risk-aware decision support in healthcare settings for improved diagnosis and patient care.

  • Reconcile data across diverse document formats

    The platform connects information across disparate files like PDFs, Excel charts, and emails to resolve contradictions and verify details. This allows for visual-to-text auditing, flagging outdated data, or verifying signed contracts against legal reviews, improving cross-document synthesis and data reconciliation.

Read the original announcement →

https://cloud.google.com/blog/topics/partners/box-ai-agents-gemini-embeddings-multimodal-enterprise-ai/

Related releases