Technology
Understand technology concepts and foundations, from AI and data to system architecture.
Secure cross-cloud infrastructure for agentic enterprises
Core principles for a unified multicloud foundation that runs AI agents and enterprise workloads securely.
AI infrastructure for systems at scale
How to prepare compute, networking, and storage for reliable AI model training and inference at scale.
Scaling compute infrastructure for enterprise applications
Expand processing capacity while maintaining performance, availability, and cost control across enterprise environments.
Sovereign cloud infrastructure for sensitive data
Explore sovereign cloud architecture, data residency, and governance requirements for regulated information.
How MCP connects AI to enterprise tools
A shared approach to connecting AI assistants with enterprise data and tools, with clear access boundaries and operational ownership.
Connecting AI answers to enterprise documents
Build grounded answers through document preparation, permission-aware retrieval, and separate checks for search and response quality.
When is a smaller model enough for an enterprise?
Assess lightweight and local models against task quality, Persian language requirements, device capacity, and maintenance effort.
What happens to tables when data structures change?
Explore Apache Iceberg schema and partition evolution, with practical checks for stable reporting and accountable operations.
Articles
Practical guides to design, evaluation and operation, focused on technical decisions and workflows.
A guide to production-ready AI agents
Move from prototype to production with evaluation, monitoring, security, and human oversight built into the lifecycle.
Secure identity and access for AI agents
Manage machine identities, permissions, and audit trails to prevent unintended access by autonomous agents.
Cybersecurity strategy for multicloud enterprises
A practical framework for unified security policies, observability, and threat response across multiple clouds.
Data architecture for enterprise AI agents
Design a governed data foundation that gives enterprise assistants reliable, current, and permission-aware information.
How do we know an AI agent did the job correctly?
Design evaluations around completed work, tool use, and the failures that matter in enterprise workflows.
When a document gives your AI assistant instructions
Understand indirect prompt injection and enforce tool permissions and trust boundaries outside the model's text response.
What does a completed AI task really cost?
Connect model, data, infrastructure, and human review costs to outcome quality and business value.
We have backups. Can we restore the service?
Turn backup completion into recovery readiness through protected copies, service priorities, and practical restoration exercises.
Latest
Recent announcements and capabilities, with sources and an explanation of their organizational relevance.
New capabilities for enterprise AI agents
A practical overview of new tools for building, orchestrating, evaluating, and deploying enterprise AI agents.
Thirteen practical demos for building AI agents
Searchable examples of AI agents for analytics, customer service, operations, and workflow automation.
Enterprise AI for legal services
Use AI for document discovery, contract review, and faster legal responses while protecting confidentiality.
Security governance for autonomous agents
Control autonomous-agent risk through policy, continuous evaluation, tool restrictions, and human oversight.
Reusable skills are coming to Gemini and Workspace
Google expands reusable instructions, offering teams a way to document working methods and make recurring outputs more consistent.
New BigQuery tools for understanding metric changes
Investigate metric changes in BigQuery with SQL examples for ML.DETECT_CHANGE_POINTS and AI.KEY_DRIVERS, plus guidance on interpreting results.
AI inference joins the Pub/Sub message stream
AI Inference SMT adds model results to event messages, creating another path for enriching data as it moves downstream.
Microsoft warns about fake passkey setup messages
A recent report describes attackers using passkey enrollment themes to deceive users and gain access to cloud accounts.