Companies are stepping up competition over platforms that optimise various AI models for internal data, helping AI understand organisations in more depth and execute tasks.

Platforms that support AI in understanding company context and serve as an operating system are being treated among enterprise software companies as a strategic battleground that could shape future hegemony.

Microsoft recently unveiled three new products at its Build 2026 conference to help AI understand corporate context. Amir Netz (아미르 네츠), CTO for Microsoft Fabric, said enterprise AI should be like an insider who knows how an organisation works. He stressed that for agents to operate reliably, they need an organisation’s memory, a "context layer."

Microsoft Fabric targets AI agent data platforms, taking aim at Snowflake and Databricks.

Cloud data platform Snowflake also used its annual Snowflake Summit to highlight a platform strategy to help AI understand corporate context and introduced a wide range of services that can use it. It also made clear its intent to expand its role beyond data storage and analytics into a gateway for companies to use AI agents.

"Competition in agentic interfaces is intensifying..." "In the end, services will follow where the data is." "An agentic OS strategy, Snowflake-style, is emerging... AI brand also to be revamped." "On the road to an agentic enterprise, watch 4 things."

Adobe also carried out major updates across Adobe GenStudio. The core of the update is newly launched Adobe Brand Intelligence. It continuously learns not only static materials such as brand guidelines but also qualitative data such as review feedback and approval and rejection histories, so AI agents can automatically produce content that matches brand identity.

Adobe adds Brand Intelligence to GenStudio, expanding the content supply chain with AI agents.

SAP also presented a strategy at the Sapphire 2026 conference held in Orlando, Florida in May that the decisive factor in the AI race is not models but corporate context and governance. SAP CTO Philipp Herzig (필립 헤르치히) stressed that large language models are not a differentiator. He said it does not matter whether a company uses an OpenAI model or an Anthropic model. He said what matters is whether an agent knows the right business entity, can access appropriate data and can be tested with real corporate data.

SAP targets an AI agent hub with an autonomous enterprise strategy.

Amid concerns over AI costs and security, the importance of the PC as the hardware that runs AI is also growing. Semiconductor companies are increasing spending on PC AI chips, and PC makers are also moving briskly. AI agents that run on PCs are also emerging one after another.

AI-driven reshaping of the PC industry becomes reality... Apple vs Nvidia big match. A Windows PC ecosystem based on Nvidia integrated chips is emerging... targeting Intel and AMD. [Tech Inside] Will people really buy new PCs because of AI?... 'Nvidia Inside' put to the test.

Other moves by companies at home and abroad around AI were also summarised.

Microsoft is moving in earnest to a strategy of its own AI models, moving away from reliance on OpenAI and Anthropic. At its annual Build developer conference, Microsoft unveiled its first reasoning model, MAI-Thinking-1, and 6 models for images, voice, transcription and coding.

Microsoft: Aim to join the "big 4" AI model developers. Microsoft joins the AI race with its own models: "We will compete on a very low price."

Microsoft unveiled a wide range of AI agent tools to support native app development for Windows 11. It also updated its platform that supports developing and operating AI agents and applications. It also unveiled a new operating system for AI agents, Project Solara, and introduced Scout, an AI agent based on the OpenClaw framework, to its cloud-based productivity platform Microsoft 365.

Microsoft steps up expansion of the Windows 11 native app ecosystem. Microsoft updates a platform supporting development and operation of AI agents and applications. Microsoft unveils Project Solara, an OS for AI agents, based on Android, not Windows. Microsoft unveils Scout, a personal AI agent based on OpenClaw.

OpenAI is strengthening its push for enterprise customers by adding 6 role-specific plugins and new features to Codex. The update focuses on expanding Codex use beyond developers to office work in general. OpenAI is also reviewing a plan to publicly release software that enables its AI to run across multiple chips.

OpenAI releases new Codex tools targeting office work. OpenAI reviews release of internal chip abstraction software to support running AI on multiple chips.

Anthropic has formalised a third-party partner programme for selling its Claude AI products, called the Claude Partner Network.

More like an enterprise tech company... Anthropic launches the Claude Partner Network.

Anthropic also urged major AI labs to slow AI development. In a blog post, it warned that AI systems are nearing a stage of "recursive self-improvement" in which they improve themselves without human intervention. It also said slowing the pace of AI development would be "beneficial to the world."

Anthropic: "We need to slow AI development"... warning that "self-improvement" is near.

Apple is reported to launch a next-generation Siri voice AI assistant in September, aiming to provide a more advanced and personalised conversation experience.

Apple to launch next-generation Siri in September... also to use Google Cloud and Nvidia chips.

Nvidia acquired enterprise predictive AI software startup Kumo AI for more than $400 million. The deal expands Nvidia’s AI model portfolio optimised for its hardware and broadens the foundation for companies to further customise its models.

"Nvidia acquires Kumo AI for $400 million"... securing AI models specialised in structured data.

Meta launched an AI agent, Meta Business Agent, on WhatsApp, Messenger and Instagram that automates customer support, product recommendations and reservation processing.

Meta launches Meta Business Agent... supporting customer responses on WhatsApp and Instagram.

Vice President Jeong-a Seo (서정아), head of the Digital Commerce Team at Samsung Electronics' MX Division, took the stage for a keynote session at the Snowflake Summit 2026 conference in San Francisco on June 2 local time. She shared the company’s ongoing AX strategy and experience in a conversation with Snowflake Senior Vice President Christian Kleinerman.

Samsung Electronics Vice President Jeong-a Seo: "The core of AX is expanding who can access data... processes also need redesign." [DT People] "AI-based SCM intelligence results are taking shape... AI-ready data is key to AX."

Ninehire, a recruiting management solution operated by Worksphere, introduced an "MCP (Model Context Protocol) integration" feature linking its solution with external AI tools. AI and cloud company MegazoneCloud built an AI shopping agent for Golfzon Commerce’s online shopping mall that provides golf equipment recommendations. NC AI won a Hanwha Ocean project titled "development of a vision recognition-based welding-only model and a collaborative-robot-based autonomous welding model" and began a collaboration system.

Ninehire releases "MCP integration"... manage recruitment directly from AI tools such as Claude. MegazoneCloud builds an AI shopping agent for Golfzon Commerce... supporting an AI fitting service. NC AI wins Hanwha Ocean autonomous welding project... to apply to merchant ships and special ships.

Global open-source solutions company Red Hat updated the Red Hat Ansible Automation Platform (AAP) to support operationalising AI agents in enterprise environments.

Red Hat upgrades Ansible Automation Platform... evolving into an operational layer for agentic AI.

A survey found that web traffic generated by AI agents has surpassed human traffic for the first time in internet history. According to Cloudflare, AI agent bots accounted for 57.4 percent of total web traffic, while human traffic stood at 42.6 percent.

AI agent web traffic overtakes humans... is the 'dead internet' theory becoming reality?