🤖 AI Agents & Automation

Autonomous AI agents, agentic workflows, tool use, computer use, and multi-agent systems.

1,798 articles

Snowflake unveils AI agent features for the 'agentic enterprise' at Expedition 2026

AI Data Cloud company Snowflake announced a major expansion of its product lineup at Snowflake Expedition 2026. The company is positioning its platform as the foundation for the agentic enterprise. Source: timesofindia.indiatimes.com

OpenAN tackles the next autonomous-network challenge: AI agents that cooperate

Telecom operators have spent years teaching networks to manage themselves. The next, far harder problem is getting multiple AI agents inside those networks to work together. OpenAN is aiming squarely at that challenge. Source: fierce-network.com

Agentic AI set to reshape agricultural research and farming

A comprehensive scientific review describes how autonomous multi-agent AI systems could transform both agricultural research and day-to-day field operations. Source: bioengineer.org

Securing AI agents: lessons from real-world deployments

Agentic AI demands a fundamentally different security model than traditional software. Unlike conventional services that run predefined logic, agents make their own decisions at runtime. Source: uber.com

Singapore weighs tighter rules for high-risk AI

Singapore is considering stricter controls on high-risk AI, particularly systems that can act autonomously and access sensitive data. Source: vietnam.vn

Five things CIOs need to know to govern Claude Code

Claude Code has quickly become one of the most widely adopted AI agents in the enterprise. Developers use it to write and refactor code and to troubleshoot builds, which raises governance questions for IT leaders. Source: cio.com

Kaspersky: don't trust LLMs by default in AI agent architectures

Kaspersky argues that enterprise security architecture should treat large language models inside autonomous AI agents as untrusted software components. Source: homeland411.com

'Web Hole' attack lures AI agents through abandoned websites

Attackers can re-register expired domains that appear in search results and poison their content. AI agents following those results could be tricked into taking harmful actions. Source: securitybrief.co.uk

Microsoft: agentic AI weaponized to steal cloud secrets within five minutes

Microsoft warns that threat actors are weaponizing autonomous agentic AI frameworks to systematically crawl open-source repositories. The goal is to harvest exposed cloud secrets, and the attacks can succeed in as little as five minutes. Source: homeland411.com

Business AI agents: who's using them, what they do and what firms spend

Drawing on research from McKinsey, Gartner and BCG, the piece examines who deploys AI agents in business. It also covers which tasks agents handle most and how much companies spend. Source: google.com

Box turns to Nvidia AI models for a context layer powering enterprise agents

Box Inc., one of the tech firms most eager to embrace AI agents, is going further down that path. It is using Nvidia's AI models to build a richer context layer for autonomous enterprise agents. Source: google.com

Firms can't tell what 'normal' behavior is for their AI agents

A new report finds that organizations are unable to establish behavioral baselines for AI agents. That gap makes it harder to secure these autonomous tools. Source: cybersecuritydive.com

Attackers wield agentic AI tooling to hunt down and exploit leaked secrets

Microsoft Threat Intelligence has seen threat actors use an agentic AI offensive framework. It scans public GitHub repositories for exposed secrets. Source: google.com

Genesis Agent Skills release aims to boost innovation across the Genesis Mission

A new public software release offers researchers plain-language access to reusable guidance and tools. The resources are meant for building, evaluating and improving Super... Source: news.fnal.gov

Perplexity launches 'Teammate' AI agent for complex work projects

Search-focused Perplexity is following Anthropic, OpenAI and Microsoft by introducing agent technology for complex workplace tasks. Source: me.pcmag.com

BMO debuts AI Agent Assist to speed up card support and refunds

BMO has rolled out AI Agent Assist, an internal agentic AI tool. It cuts credit card inquiry resolution times by 30% and automates 40% of refund handling. Source: google.com

Merge builds product and business tools on OpenAI technology

Merge uses GPT-6 Astra, ChatGPT and Codex to review product code and support marketing. Its teams also turn to AI for planning executive dinners and preparing sales materials. Source: google.com

Containing autonomous AI agents with rootless Docker

The article explains how rootless Docker, Linux user namespaces and cgroups v2 confine autonomous AI coding agents. The setup prevents them from escalating to root privileges on the host. Source: google.com

AWS wants developers to stop babysitting their AI agents

Developers were promised that AI agents would lighten their workload, yet running several agents in parallel has turned into a job of its own. AWS is aiming to cut that supervision burden. Source: theaieconomy.substack.com

China's AI agents hit a wall

Earlier this year, footage of retirees and young people queuing for hours to install the OpenClaw AI agent went viral in China. The enthusiasm, however, appears to be running into obstacles despite government support. Source: bloomberg.com