Healthcare Practitioners and Technical Workers, All Other

A VentureBeat Pulse Research report reveals that enterprise AI organizations face a deployment problem, not a platform problem, with most deployed 'agents' still being basic chatbots despite ambitions for multi-step orchestrated workflows. Companies are consolidating on model-provider platforms like Anthropic, driven by 'model gravity,' but plan hybrid control planes to avoid vendor lock-in. Investment flows to agent workflow tooling and security, yet real-time fiscal control over token consumption lags.

A new "agentjacking" vulnerability allows attackers to inject malicious code into AI coding agents via trusted error reporting tools like Sentry, bypassing traditional security, and necessitating urgent workflow changes for security and development teams regarding AI agent governance, runtime detection, and access controls. This systemic flaw impacts how enterprises manage AI agent security at an ecosystem level, requiring significant shifts in industry practices.

OpenAI and Molecule.one show how a near-autonomous AI chemist using GPT-5.4 improved a key drug-making reaction, advancing medicinal chemistry research.

Moonshot AI released Kimi K2.7-Code, an update to its coding model claiming 30% reduction in thinking-token usage and improved code generation in Rust, Go, and Python, impacting inference costs for teams running agentic workflows and production gateways, particularly in frontend development, DevOps, and performance optimization.

Microsoft and Mayo Clinic are partnering to create an advanced AI model tailored for the healthcare sector, aiming to introduce new capabilities for healthcare practitioners and technical workers.

When someone on a team corrects an AI agent — better prompts, better feedback, better context — that improvement disappears the moment a colleague opens the same tool. The correction doesn't transfer, and the next person starts from zero. The problem compounds in multi-agent workflows, where teams expect agents to share context across users and tasks. Without a shared memory layer, every team member effectively trains a different version of the same agent — and those versions never sync. That gap shows up in the numbers. According to Asana's own research, 75% of knowledge workers use AI on the job, but only 5% of companies have reported productivity gains. “Model providers are getting really, really good at improving reasoning and retry loops, but what they’re not good at is bringing the enterprise work context in a way that human beings can reason about for shared memory,” Asana Chief Product Officer Arnab Bose told VentureBeat. Asana had been building toward an agentic platform that centers context and shared memory. Its Agentic Work Management platform ensures that if any team member corrects an agent, that correction applies to everyone else on the team. “That context graph is automatically provided to agents operating inside Asana’s system so you don’t have to have every human member of the team become an expert at prompt engineering or context engineering,” Bose said. Bose said the shared memory architecture matters beyond Asana's own product; it's the design decision enterprises need to make for any multi-agent system. Shared memory also becomes important when enterprises begin moving from simple single agents to multi-agent workflows that need to share context and behaviors. Memories for a multi-agent, multi-platform workflow The models powering agents are stateless by design, so memory becomes a dedicated layer outside of a context window. While this area of AI innovation is marching towards maturity, the question of what gets stored, who controls it, and how it stays consistent when different agents and users write to the same instance remains largely unsolved. This is manageable for use cases with only one user. However, in enterprise agentic workflows, the idea is for agents to work with the entire team. Most platforms have agents that still act for individuals, which leads to task repeating and inconsistent versions of reality and spreading mistakes. Agents could then also contradict each other. Sriharsha Chintalapani, co-founder and CTO of Collate, said in an email to VentureBeat that the lack of shared memory is a major obstacle for multi-agent workflows particularly around consistency. "Agents are sensitive to the quality of their prompts," Chintalapani said. "Someone with a strong understanding of the task will generally get more accurate results than someone less experienced. Partly that’s because they’re able to construct more detailed prompts, but also because they’re able to give the agent better feedback. The agent remembers the corrections it’s received and applies that knowledge to successive prompts. The more accurate the feedback, the better the agent will perform for that user. " He added that organizations should stop treating shared memory solely as a prompt engineering problem and think of building systems that repeat context across every conversation. Neej Gore, chief data officer at Zeta Global, said in a separate email that shared context becomes a living memory that "compounds intelligence across the enterprise." The opportunity may lie in building AI agents that retrieve memory relationally, pulling in relevant context based on what’s being asked — an approach Chintalapani says few organizations outside the largest model providers are equipped to build. Personal versus team agents AI agents already proliferate enterprises; it’s just that many of these operate as personal agents doing work specific to individual users. Most prompts start from one person, any files are uploaded by one account, and even for agents living in a company-wide system mostly learn individual user preferences. Most enterprise AI workflow platforms recognize that memory is important but approach it through different lenses. For example, Microsoft’s Copilot takes an individual-first approach by learning a user’s role within the organization, tone preferences and working patterns, which are then stored as personal memories for the agent to apply across the different Microsoft 365 surfaces. For engineering and orchestration teams evaluating agentic platforms, the shared memory question is now a procurement criterion — not just a technical nicety. An agent that learns only for the person using it will require ongoing individual upkeep. One connected to a team-wide memory layer builds institutional knowledge automatically.

Google launched Gemma 4 12B, an open-source multimodal AI model designed for local execution on enterprise laptops with 16GB VRAM, featuring an encoder-free architecture for audio and video, a 256K token context window, and native agentic tool-use capabilities, targeting privacy-sensitive and edge AI deployments.

GitHub published a guide for developers on how to effectively review pull requests generated by AI agents, focusing on identifying issues and managing technical debt.

OpenAI is providing ChatGPT for Clinicians free to verified U.S. physicians, nurse practitioners, and pharmacists to support clinical care, documentation, and research workflows.

A VentureBeat survey reveals that most enterprises are ill-equipped to handle stage-three AI agent threats, citing incidents like data exposure at Meta and a supply-chain breach at Mercor. The survey highlights a common security architecture gap: monitoring without enforcement, and enforcement without isolation. Executives often overestimate their protection, with 88% reporting AI agent security incidents in the last year, but only 21% having runtime visibility. The article outlines an AI agent security maturity audit with three stages (Observe, Enforce, Isolate) and a 90-day remediation sequence, detailing attack scenarios, detection tests, blast radius, and recommended controls. It emphasizes the need for scoped agent identity, approval workflows for write operations, and sandboxed execution, noting that current hyperscaler offerings and open-source frameworks often lack complete stage-three capabilities. CISOs and security leaders are urged to move beyond basic monitoring to implement robust enforcement and isolation strategies to mitigate increasing machine-speed threats and regulatory risks.