Google Adds Agentic Workflow Steps to Opal No-Code App Platform

In Brief

• Opal now features agent-driven workflow steps for dynamic task handling.
• Agents gain memory, dynamic routing and interactive chat for tailored experiences.
• Workflow creators can automate more complex, personalized processes with less manual setup.

Google upgraded its Opal workflow platform from static model calls to agentic intelligence, allowing AI to determine the best path to complete complex tasks.

Users can now select an agent in the "generate" step that proactively triggers appropriate tools and models based on stated goals. The update introduces three new capabilities: persistent memory across sessions, dynamic routing based on custom logic and interactive chat for gathering user input.

Google claims these features will make workflows feel more like collaborative partners than one-way processes.

Google introduced the Opal "experiment" to the U.S. in July 2025 as a no-code, natural language platform to create lightweight apps. Google uses the experiment language for initiatives it hasn't fully committed to, so users shouldn't have any expectations of continuity of features or of the platform itself.  

The company introduced new debugging features and expanded access to the platform in October 2025, opening it up to 15 more countries. 

New Opal Agent Capabilities

Google states the update provides workflow builders with the power of automation while maintaining step-by-step control.

Capability	Description
Agent Step	Selects tools and models based on user goals automatically
Memory	Remembers user preferences and information across sessions
Dynamic Routing	Enables multiple workflow paths based on custom logic
Interactive Chat	Initiates dialogue to gather missing information or offer choices
Tool Integration	Connects to Web Search, Veo for Video and other models

Google's Enterprise AI Play

Google intensified its enterprise AI strategy throughout 2025-2026, positioning Workspace as a direct competitor to Microsoft. In April 2025, Google opened Workspace to third-party AI models like Anthropic's Claude, while reporting that Gemini delivers over 2 billion AI assists to business users monthly.

At Google I/O in May 2025, the company unveiled 100 AI-focused announcements integrating Gemini across Docs, Sheets, Slides and Gmail, and introduced real-time translation in Google Meet.

The company introduced Google Enterprise in October, a platform that brings together its AI models, productivity tools and connections to major business software like Salesforce and SAP into one unified system — allowing companies to automate complex, multi-step tasks across their entire workplace rather than relying on disconnected AI tools.

In January 2026, Google launched the Feeds app for Google Chat, consolidating external content streams within Workspace. In February 2026, Google enabled cross-platform meeting interoperability between Google Meet and Microsoft Teams devices.

Have a tip to share with our editorial team? Drop us a line: