Large Language Models

Today's Stories

1. 1

   Cisco deploys AI agents to all 90,000 employees

   Cisco is rolling out personalized AI assistants to its approximately 90,000 employees starting in its new fiscal year at the end of July. Each agent can handle tasks, answer questions, and route requests to the most efficient AI model; the system dynamically selects the right tool rather than relying solely on frontier models. The rollout reflects how large enterprises are embedding AI into core operations, not just infrastructure. CFO Mark Patterson noted that AI is "the most significant technology transition" Cisco has seen. On the finance side alone, AI now handles 80–90% of the first draft of MD&A preparation (mandatory narrative sections in public company filings), and Cisco is building AI tools for investor relations and a CFO dashboard to synthesize performance data and predict business direction.

   Patterson expects internal competition as teams discover new AI applications, supported by company-wide upskilling and knowledge-sharing efforts. Cisco has also signaled major growth from hyperscalers (large cloud providers): the company reported $2 billion（約3200億円） in orders in fiscal year 2025 and has raised its fiscal year 2026 guidance to $9 billion（約1.4兆円）.

2. 2

   Accenture, ServiceNow Partner on AI-Powered Security Migration Services

   Accenture and ServiceNow announced a partnership to build managed security services that combine ServiceNow's AI platform with Accenture's AI capabilities. The services are designed to help clients migrate from legacy security tools, manage third-party and operational technology risks, accelerate compliance, and deliver faster, outcome-based cybersecurity services. Many enterprises remain stuck on decades-old security technologies and struggle to keep pace with rising AI-driven threats. By automating what were previously manual, labor-intensive security and compliance tasks—such as tracking vendor supply chains and managing European AI regulations—the partnership allows organizations to reduce compliance time potentially by half or more, while Accenture assumes some of the migration and operational risk.

   Accenture already has a couple of clients on the new platform and reports a significant and growing pipeline. The service bundles existing ServiceNow products (including technology from its acquisitions of Armis and Veza) alongside Accenture's proprietary migration and AI services, delivered as an end-to-end outcome-based model rather than traditional consulting or time-and-materials engagement.

3. 3

   Joint Research Using Real-World Data and Generative AI (LLM) to Improve Accuracy and Efficiency in Clinical Trial Candidate Patient Extraction

   Joint Research Using Real-World Data and Generative AI (LLM) to Improve Accuracy and Efficiency in Clinical Trial Candidate Patient Extraction

4. 4

   GPU Processing Performance 475 Times Faster...Fujitsu Introduces New Design Concept for LLM

   GPU Processing Performance 475 Times Faster...Fujitsu Introduces New Design Concept for LLM

5. 5

   Kindai University Hospital, Nakaitai Seiyaku, NTT and Others to Extract Clinical Trial Candidate Patients Using Domestically-Developed LLM, Four Organizations Launch Joint Research

   Kindai University Hospital, Nakaitai Seiyaku, NTT and Others to Extract Clinical Trial Candidate Patients Using Domestically-Developed LLM, Four Organizations Launch Joint Research

6. 6

   OpenAI develops custom AI chip for text generation with Broadcom

   OpenAI has developed a custom AI chip designed specifically for inference—the process where an AI model produces answers—in collaboration with Broadcom, a chip design company. Building its own chip gives OpenAI more control over the hardware that runs its AI models, potentially reducing reliance on external chip suppliers and lowering costs for the company's operations.

   The article does not specify a release date, availability timeline, or pricing for this custom chip.