# AI agent use cases

An agent becomes more useful when its work survives the chat. These examples
turn common jobs into persistent, searchable datasets without asking you to
design the backend from scratch. Each guide includes a starter schema, stable
index, agent jobs, and the Rowset features that matter.

Start here when you know the job but do not want to design a dataset from scratch.

## Choose a workflow

- [Personal CRM](/use-cases/personal-crm)
- [Agent task board](/use-cases/agent-task-board)
- [Feedback triage](/use-cases/feedback-triage)
- [Content pipeline](/use-cases/content-pipeline)
- [Product or inventory catalog](/use-cases/product-inventory-catalog)
- [Bug or QA tracker](/use-cases/bug-qa-tracker)

## How to choose

Use a use-case page when you need a concrete dataset shape. Use
[Create datasets](/docs/create-datasets) when you already know the shape. Use
[Design dataset schema](/docs/design-schema) when the main question is column
types, status rules, references, or instructions.

## What reliable agent workflows have in common

- a stable index column
- small, clear headers
- dataset instructions that explain workflow rules
- semantic column types for dates, URLs, emails, choices, money, images, audio, or references
- optional project grouping when several datasets belong together
- optional public preview only when a human needs read-only access