Box integrates Gemini Multimodal Embeddings 2 into its Agentic Platform to extend retrieval-and-agent workflows beyond text, preserving table structure, page layout, and visual content across mixed file types.
Multimodal embeddings embed text, images, rendered pages, tables, and charts into a single semantic space so agents can retrieve visual and spatial elements as easily as prose.
Preserving layout and table semantics prevents loss of meaning that happens when converting structured data to plain text, enabling accurate financial and analytical queries.
Cross-format retrieval lets an agent link a PDF policy page, a spreadsheet table, and a presentation slide without manual tagging or format-specific workarounds.
# What changed
# Why it matters
# Core capabilities introduced
# How this changes agent behavior
# Three design patterns where multimodal agents add value 1) Complex financial and analytical reporting The embedding of rendered tables and charts preserves row/column relationships so agents can surface exact supporting tables or flag discrepancies between written claims and visual trends.
2) Multimodal clinical decision support
3) Cross-format operational workflows
# Practical effects for teams
# What to watch next Adoption will depend on how teams pipeline existing Box content into layout-aware embeddings and how agents are trained to interpret those multimodal vectors for tasks like trend detection, anomaly identification, and document-level sourcing.
# Short summary Box plus Gemini Multimodal Embeddings 2 moves enterprise agents beyond text by embedding images, pages, tables, and charts into a unified vector space. That preserves spatial and structural meaning across mixed file types and enables agents to retrieve and reason over the precise visual or structured element that supports a query.
Users can now securely browse, access, and work with Box content directly in ChatGPT, bringing enterprise knowledge into AI-powered workflows Box, Inc., the leading Intelligent Content Management (ICM) platform, announced it is bringing enterprise content and AI together to power more intelligent, actionable ways of wo
Few professions are as exacting as the practice of law. A team reviewing a contract or building a case works inside strictly privileged information, firm-specific playbooks, and a body of law that changes constantly. The work thrives on nuanced, professional judgment — and the systems supporting it inherit real obligat
Protecting capital in today's markets requires immense speed and precision. A financial analyst preparing a deal memo works across licensed market data, internal models, and confidential client files. General-purpose AI lacks the real-time accuracy, verifiable data lineage, and strict security that financial institutio
We’re launching agentic video understanding across our latest Gemini models for improved accuracy and lower costs and token usage.
We are excited to share that Gartner has named Google a Leader in its inaugural 2026 Magic Quadrant for Enterprise AI Assistants . In this comprehensive evaluation of top enterprise AI assistant vendors, Gartner placed Google in the Leaders quadrant for its evaluation across both Completeness of Vision and Ability to E
Join Google's SMB experts for a hands-on look at three practical Gemini Enterprise use cases you can set up in 10 minutes or less.
Loading more related stories...
Open the app view to save this story, compare related coverage, and continue from the same source.