AI Agents

Akamai released a security survey for the agentic AI era, highlighting the risk of sensitive information leakage in chatbot conversations.

Introduces a case study identifying why team productivity remains at 1.1x despite 100% AI agent adoption, using a three-factor decomposition analysis.

• •Yasuhiro Okumura of NEC Solution Innovators faced the challenge of overall team productivity remaining at only 1.1x.
• •Team speed was analyzed by decomposing it into three factors: 'individual speed', 'scope where speed is effective', and 'flow efficiency'.
• •Efforts to resolve bottlenecks were discussed in a session at Developers Summit 2026 KANSAI.

Meta's new enterprise AI business, Meta Enterprise Platform, began offering AI models and agent swarms, with a former MongoDB CEO appointed as head.

• •It provides AI models and agent swarms to enterprises and developers.
• •Chirantan Desai, who immediately resigned as President and CEO of MongoDB, was appointed as the head.
• •It incorporates security and plans to offer dedicated VMs that even Meta cannot access within the year.

US-based GitLab released GitLab 19.4, the latest version of its Git repository manager, enhancing support for agentic automation.

• •US-based GitLab announced the release of the latest version, GitLab 19.4, on the 28th.
• •This is a version upgrade of the Git repository manager GitLab.
• •A key feature is the enhanced support for agentic automation.

Explains front-line development practices like harnesses and specification-driven development in response to AI agent adoption challenges.

• •While AI agents are changing software development, adoption challenges exist, such as inconsistencies with existing codebases and token consumption.
• •Three speakers from KDDI Agile Development Center explained AI agent classifications and trends in custom-built agents.
• •Introduced cutting-edge front-line concepts: harnesses, specification-driven development (SDD), and loop engineering.

This article explains how OEDG, four years into its DX, is challenging physical AI, overcoming regulatory industry constraints to foster a 'try first' culture.

• •After being criticized by outside directors for lack of digital engagement, OEDG moved into physical AI using drones and robots.
• •The focus is on how they built a culture of trial and error in a regulated industry where failure is difficult to tolerate.
• •An interview with Masao Koizumi, who leads the company's DX, details specific initiatives and the background of organizational change.

JetBrains announced JetBrains Air, a new initiative that supports multi-vendor development using AI agents.

LM Studio added a collaborative editing canvas to its AI agent 'Bionic,' enabling direct implementation from blueprints.

• •LM Studio announced on September 23 that it added new features to its PC-based AI agent 'Bionic.'
• •The added feature is a canvas that allows users and AI to edit collaboratively.
• •This feature enables direct code implementation from blueprints or designs.

The article explains the importance of "harness engineering" for AI agents and analyzes the current status and future strategies of Japanese companies.

• •A trend is shifting from AI model performance debates to "harness engineering" for autonomous operation environments.
• •It is pointed out that Japanese companies' efforts in this field are "almost non-existent."
• •Hiroaki Kaminomoto of ITR explains realistic solutions for companies to remain competitive.

Shopify has extended WebMCP support to its checkout feature, enabling browser-based AI agents to update order details and complete purchases with buyer approval.

• •Shopify has expanded the scope of its WebMCP support to include the checkout feature.
• •Browser-based AI agents can now update order details and complete purchases.
• •These operations are performed only with the buyer's approval.