Prompts Overview

Prompts are reusable, versioned AI instructions - you build one once, configure the model and settings, and anyone on your team can run it from chat, a Project, a Task, or an API call. For where prompts sit among Aisle's other building blocks, see How Aisle Works.

The Building Your First Prompt guide walks through creating a prompt from scratch, configuring every setting, and deploying it to your team.

When to reach for a prompt

Use a prompt when the same AI task comes up repeatedly and you want consistent, shareable results:

• Direct chat - one-off conversations, exploring ideas, trying different approaches.
• Prompts - repeatable tasks with standardized instructions and consistent outputs, shared across your team.
• Tasks - deterministic Python automations that combine prompts with logic, integrations, and scheduled or triggered runs.

Prompts sit in the middle - more structured than free-form chat, lighter than a full Task.

Common use cases

• Content translation that maintains your brand voice
• Customer feedback analysis that identifies themes your business cares about
• Document formatting that adapts your writing for different audiences
• Data extraction from unstructured text
• Code review with your team's standards
• Meeting notes summarization in your format

Next steps

• Building Prompts - create and configure a prompt: instructions, variables, model settings, tools, and structured output
• Deploying Prompts - share it, version it, expose an API endpoint, and review run logs