Enterprises deployed AI agents knowing the controls weren't ready, new VentureBeat Research finds. Now 54% have had a security incident, and 27% only find out what an agent costs when the invoice arrives. Full reports drop at VB Transform, July 14–15 in Menlo Park. Read now → https://bit.ly/4vZydCI

Enterprise AI agent controls lag deployment | VentureBeat

VentureBeat Research surveyed 573 enterprise leaders and found AI agent controls lagging deployment, with 86% of GPUs running at half capacity or less.

bit.ly

4 days. That’s all the time left between you and VB Transform. 🚀 No fluff, no high-level hype just real builders sharing how they're shipping AI to production and solving architecture bottlenecks. Grab one of the final spots here: https://bit.ly/4f4Jvhw

Only 5 days left until VB Transform! If you haven't locked in your spot yet, this is your sign. We’re bringing together the brightest minds to talk about what’s next in the industry. Don't get left behind. Secure your spot now 👇 https://bit.ly/4eX8y6h #VBTransform2026 #Innovation #Networking

Next week, our very own Louis Columbus is heading to #VBTransform! Catch his insights on the enterprise AI trust gap including why 54% of leaders have already faced an AI agent security incident and connect with the team. Meet him on July 14-15 in Menlo Park. Register today: https://bit.ly/4f4Jvhw

8 days out. The gap between "works in the demo" and "runs unsupervised without falling over" that's basically what VB Transform is about this year. Passes are limited. Request your invitation: https://bit.ly/4eX8y6h July 14–15 | Hotel Nia, Menlo Park

Trunk Tools built a three-layer architecture (perception, semantics, agents) based on highly-detailed, industry-specific data. It's translatable across all verticals. https://bit.ly/4y4eenC

Trunk Tools ditches general-purpose models | VentureBeat

General-purpose models struggle with messy, industry-specific data. A three-layer AI stack from Trunk Tools cut document review cycles from 60 days to 10.

venturebeat.com

Prompt injection techniques have evolved, and cybercriminals are taking advantage of the disconnect between assumptions about #LLMs and their actual characteristics. #AI #ArtificialIntelligence #Security https://bit.ly/44wm8Zw

Prompt injection is exploiting enterprise AI's biggest design flaws by targeting agents, RAG pipelines and model routers

In the past two years, businesses have been trying to fit large language models (LLMs) into support, analytics, development, and internal automation like never before.

venturebeat.com

Thanks to Claude Code, Anthropic's engineering team is shipping at roughly three times its actual headcount. The bottleneck has moved from the IDE; devs have to decide what to actually build. #AI #ArtificialIntelligence #GenerativeAI https://bit.ly/4oQLxqj

Claude Code turned every engineer into three. Now companies need more product thinkers

AI compressed the build. Fundamentals matter more, not less, and the product funnel is now where engineers earn their keep.

bit.ly

Morgan Stanley deployed #AI agents in one of banking's most accuracy-critical, deadline-driven workflows — and cut the work in half. The trick: Making agents less autonomous, not more. #GenerativeAI https://bit.ly/4wicYM3

Morgan Stanley cuts P&L reconciliation in half | VentureBeat

How Morgan Stanley got a regulated, deadline-driven workflow comfortable with agents: process mapping first, fixed rules, human sign-off on every call.

venturebeat.com

13 days until VB Transform 2026. The Pilot Era is Over. The Agentic AI Era is Here. Real agentic AI projects. Proprietary research. Zero vendor fluff. July 14–15 | Hotel Nia, Menlo Park Seats are going: https://bit.ly/4eX8y6h #VBTransform2026

Autonomous agents are being attacked by autonomous threats. Defenses aren't keeping up. Visa's Rajat Taneja shares their technical playbook at #VBTransform2026, July 15 — including an open-source vulnerability pipeline they've released to the community. Read more → https://bit.ly/4xR9GB3

95% of drug discovery projects fail due to handoff failures between teams. Stanford deployed thousands of AI scientist agents in a unified pipeline to fix that. James Zou presents the architecture at #VBTransform2026, July 15 in Menlo Park. Read more about his session → https://bit.ly/4ahSpXD

Intuit scrapped its multi-agent setup and rebuilt from scratch. Skill-and-tool architecture. Humans in the loop. Brain separated from execution. Nhung Ho (VP of AI, Intuit) breaks down every decision at #VBTransform2026 — July 14–15, Menlo Park. Read more: https://bit.ly/4w8PMQo

Intuit will show off how it rebuilt its AI infrastructure to support fast and complex tasks at VB Transform 2026

Customer expectations have shifted from simple, fast conversational interactions to complex agentic AI-powered tasks that legacy IT architectures simply can’t handle.

venturebeat.com

Final call — VentureBeat AI Impact Series NYC is 3 days out and reviewing last invitation requests. Building enterprise agentic infrastructure and want to be in a room with fewer than 100 others doing the same: https://bit.ly/3PXwiPr

Agentic Alpha: Engineering High-Velocity Autonomy in a Millisecond Economy · Luma

By the second half of 2026, the "Agentic Gap" has become the defining divide between the industry leaders and the followers. In the high-velocity markets of…

bit.ly

5 weeks out. Seats are moving. 🚨 VB Transform 2026 brings together the architects and engineering leaders building multi-agent systems at scale — real case studies, zero vendor fluff. July 14-15 | Menlo Park Reserve your spot: https://bit.ly/4v2MOg4

ICYMI: VB Transform 2026 covers the full agentic stack — orchestration, context layers, observability, inference, security. Speakers from Visa, Target, Mastercard, Waymo, LangChain. Proprietary research. Zero vendor fluff. July 14–15 | Menlo Park https://bit.ly/4v2MOg4

Enterprise agentic AI, same story: impressive pilot, clean demo, strong benchmarks — then it fails to scale. The failure mode? Orchestration was bolted on, not designed. That's the room. June 24, NYC. VentureBeat AI Impact Series: https://bit.ly/3PXwiPr

The "Agentic Gap" is no longer theoretical, and the enterprises winning are building multi-agent ecosystems that execute at 100x manual speed. June 24: VentureBeat's AI Impact Series returns to NYC. Fewer than 100 seats. Invite-only. Request your invitation: https://bit.ly/3PXwiPr

Agentic AI is reshaping work—not by replacing us, but by enhancing how we think and decide. Imran Aftab shares in @venturebeat.com how we can evolve with #AI: build fluency, rethink workflows, and strengthen oversight. A smarter, human-centered future starts now. Read article: bit.ly/43tOgvk

Great nuanced take on Demis vs Sergey approaches toward #AGI #ArtificialIntelligence @Google $GOOG thanks to @venturebeat.com

At Google I/O, Sergey Brin makes surprise appearance — and declares Google will build the first AGI

Google co-founder Sergey Brin makes surprise appearance at Google I/O, declaring "Gemini will be the very first AGI," revealing philosophical tensions with DeepMind CEO Hassabis who urges scientific c...

venturebeat.com

The whale has returned! 🐳 venturebeat.com/ai/deepseek-...

DeepSeek R1-0528 arrives in powerful open source challenge to OpenAI o3 and Google Gemini 2.5 Pro

Additionally, the model’s hallucination rate has been reduced, contributing to more reliable and consistent output.

venturebeat.com

Astronomer’s $93M raise underscores a new reality: Orchestration is king in AI infrastructure venturebeat.com/ai/astronome...

Astronomer’s $93M raise underscores a new reality: Orchestration is king in AI infrastructure

Astronomer secures $93 million in Series D funding to solve the AI implementation gap through data orchestration, helping enterprises streamline complex workflows and operationalize AI initiatives at ...

venturebeat.com

Qwen swings for a double with 2.5-Omni-3B model that runs on consumer PCs, laptops venturebeat.com/ai/qwen-swin...

Qwen swings for a double with 2.5-Omni-3B model that runs on consumer PCs, laptops

The Qwen2.5-Omni-3B model is licensed for non-commercial use only under Alibaba Cloud’s Qwen Research License Agreement.

venturebeat.com

OpenAI rolls back ChatGPT’s sycophancy and explains what went wrong venturebeat.com/ai/openai-ro...

OpenAI rolls back ChatGPT’s sycophancy and explains what went wrong

Many organizations may also begin shifting toward open-source alternatives that they can host and tune themselves.

venturebeat.com

The ‘era of experience’ will unleash self-learning AI agents across the web—here’s how to prepare venturebeat.com/ai/the-era-o...

The ‘era of experience’ will unleash self-learning AI agents across the web—here’s how to prepare

AI visionaries predict an 'Era of Experience' where AI learns autonomously, and it will have important implications for application design.

venturebeat.com

From friction to flow: Why Swissport scrapped its VPN maze for Cato’s SASE fabric venturebeat.com/security/swi... #CatoNetworks, #DigitalTransformation #NetworkSecurity #SASE #SecureAccess #Swissport #ZeroTrust

From friction to flow: Why Swissport scrapped its VPN maze for Cato’s SASE fabric

Swissport ditches legacy tech, deploying a global, Zero Trust SASE architecture with Cato Networks securing 26,000 users, unlocking real-time control.

venturebeat.com

#RSAC2025 Starter Pack. Red Bull? check. Party invites? check. Best booth giveaways including whack-the-attack? check. #cybersecurity #humor

Hugging Face submits open-source blueprint, challenging Big Tech in White House AI policy fight venturebeat.com/ai/hugging-f...

Hugging Face submits open-source blueprint, challenging Big Tech in White House AI policy fight

Hugging Face challenges Big Tech in White House AI Action Plan submission, arguing open-source models match commercial performance while democratizing access and enhancing national security.

venturebeat.com

OpenAI’s new voice AI model gpt-4o-transcribe lets you add speech to your existing text apps in seconds venturebeat.com/ai/openais-n...

OpenAI’s new voice AI model gpt-4o-transcribe lets you add speech to your existing text apps in seconds

Three, all new proprietary voice models called gpt-4o-transcribe, gpt-4o-mini-transcribe and gpt-4o-mini-tts.

venturebeat.com

Anthropic just gave Claude a superpower: real-time web search. Here’s why it changes everything venturebeat.com/ai/anthropic...

Anthropic just gave Claude a superpower: real-time web search. Here’s why it changes everything

Anthropic launches real-time web search for Claude AI, challenging ChatGPT's dominance while securing $3.5 billion in funding at a $61.5 billion valuation.

venturebeat.com