This Past Week in AI

Edition #1·February 28, 2026·Week of Feb 22–28

This week's AI landscape was dominated by one theme: agents moving from concept to production. From enterprise platforms to developer toolchains, the industry is rapidly building the infrastructure for AI systems that don't just suggest — they execute.

Figma Integrates OpenAI Codex, Bridging Design and Code

Figma announced a deep integration with OpenAI's Codex that lets teams move seamlessly between visual design and coding environments. Using Figma's Model Context Protocol (MCP) server, designers and developers can iterate on designs within code workflows or refine code-generated assets visually — without switching tools. For software consultancies, this signals that the design-to-development handoff is becoming a continuous loop rather than a sequential process. Source

Anthropic Launches Enterprise Plugins for Claude

Anthropic introduced customizable plugins that let Claude act directly inside enterprise tools — Excel, PowerPoint, Google Drive, Gmail, and other connected systems. Instead of returning instructions, Claude can now complete multi-step actions autonomously once prompted. Power users can design and train plugins for specific business units, positioning Claude as a central AI operating layer for enterprise workflows. Source

Red Hat Launches "Metal-to-Agent" AI Enterprise Stack

Red Hat introduced Red Hat AI Enterprise, a unified platform for deploying AI models, agents, and applications across data centers and hybrid clouds. Built on OpenShift and RHEL, the stack bundles high-performance inference via vLLM, model tuning, agent deployment, and observability — with hardware support spanning NVIDIA Blackwell Ultra, AMD MI325X, and CPU-based inference. For enterprises moving beyond AI pilots, this offers a governed, repeatable path to production-grade agentic AI at scale. Source

Snowflake Extends AI Coding Agent Beyond Its Own Ecosystem

Snowflake expanded its Cortex Code CLI to support dbt and Apache Airflow workflows, moving beyond Snowflake-native pipelines. Developers now get secure, context-aware AI assistance for model development, debugging, and optimization inside their preferred tools. This is a meaningful step toward standardizing AI-assisted development across the full data lifecycle — particularly relevant for data engineering teams working across multiple platforms. Source

Confluent Connects Streaming Data to AI Agents via A2A Protocol

Confluent expanded Confluent Intelligence with new Streaming Agents that use the Agent2Agent (A2A) protocol to trigger and coordinate external AI agents directly from real-time Kafka streams. The integration bridges Anthropic's MCP, agent frameworks like LangChain, lakehouses (Databricks, Snowflake), and SaaS platforms (Salesforce, ServiceNow). For teams building event-driven architectures, this turns streaming analytics into automated, agent-driven actions. Source

Agentic AI Foundation Surpasses 97 Members in First Quarter

The Agentic AI Foundation (AAIF), housed in the Linux Foundation, reported 97 Gold and Silver members in its first quarter — more than doubling early CNCF membership growth. Key milestones include MCP Apps (the first official Model Context Protocol extension adding interactive UIs) and deepening collaboration between OpenAI and Anthropic on shared open standards. Seven working groups are now active covering identity, trust, observability, traceability, and governance. Source

ServiceNow Launches Autonomous Workforce for Enterprise AI

ServiceNow unveiled its Autonomous Workforce — AI specialists designed to execute enterprise work end-to-end with built-in governance and operational rigor. The first specialist, a Level 1 Service Desk AI, reportedly handles over 90% of employee IT requests at ServiceNow itself with 99% faster resolution than human agents. Paired with EmployeeWorks (integrating Moveworks' conversational AI), this signals the enterprise software industry's pivot from AI features to AI execution platforms. Source

From Our Perspective

The throughline this week is clear: AI is moving from assistant to autonomous operator. Whether it's Figma bridging design and code, Confluent connecting streaming data to AI agents, or ServiceNow deploying AI that resolves tickets end-to-end — the pattern is the same. The organizations that will benefit most are those building the software infrastructure to support these new capabilities.

At Accelerate Data, we help teams architect and build the custom software, AI integrations, and data pipelines that turn these industry shifts into competitive advantages. If your team is exploring how to move from AI experimentation to production-grade implementation, let's talk.