This week's release introduces new model support, a streamlined user interface and deeper functionality in both Prompt and Workflow tooling. These features improve how teams build and operate AI processes across exploratory, operational, and agentic use cases.
New Models Available: Claude, Gemini, Grok
We have added support for the latest LLMs:
- Claude 4 (Sonnet & Opus)
- Gemini 2.5
- Grok 3
To enable these models for use across your organization, visit: Settings → Models.
These models are fully supported in multi-model chat, reusable prompts, and execution within workflow nodes.
Web Search in Prompts and Workflows
Prompts in Aisle now support the Web Search capability - just like chat. You can enable these with a simple toggle for compatible models like GPT-4.1. When turned on, these prompts automatically retrieve real-time information whether used in chat, or as part of a workflow.
This unlocks an entirely new class of dynamic, data-aware workflows:
- Run loops over trending news or search results
- Trigger LLM decisions based on up-to-date info
- Build agents that summarize web content or surface competitive intel in real time
Simplified Prompt Creation
We've refined the prompt-building workflow to help you move from idea to functional component with less friction. One powerful use case: building consistent personality prompts for internal assistants, customer-onboarding agents, or internal support tools. Versioning, structure, and LLM targeting ensure performance stays thoughtful and on-brand.
Workflow Run Mode: Now in the Builder
You can now test and execute workflows directly inside the visual editor.
This allows for rapid iteration and validation of logic without breaking flow or navigating away from the canvas.