Resources
Glossary
The words these docs use, in plain terms.
Terms
| Term | Meaning |
|---|---|
| AI agent | An AI model that decides the steps a request needs and takes them with tools. Sands is an AI agent. |
| AI teammate | How Sands works with you: you delegate work in plain language and it reports back with results and evidence. |
| Approval mode | Your setting for when Sands asks before changing your tools: Manually approve, Automatically approve or Skip all approvals. |
| Approval card | The card in the chat that shows a change before it runs, with Approve and Decline. |
| Artifact | Something a conversation produces that you keep using: a plan, a report, a file. |
| Brief | What you hand Sands to plan from: a message, a client brief, a statement of work. |
| Connection | Sands' access to one of your tools, granted through the tool's sign-in. |
| Credit | The unit Sands uses to measure work. Each task uses the credits its work took. |
| Open question | Something a plan needs that the brief did not say. Listed in the plan for you to answer. |
| Plan | The document Sands drafts from a brief: goal, capacity, tickets, risks, dates and open questions. Also called a spec. |
| Run | The steps Sands takes to answer one request, shown at the top of the reply. |
| Send | Creating a plan's work in Jira, Linear or Asana. A one-time handoff. |
| Step | One unit of work in a run, like reading a project or creating a ticket, with what it found. |
| Ticket | One piece of work in a plan. Becomes a story, issue or task when sent. |
| Workstream | A group of related tickets in a plan. Becomes an epic, milestone or section when sent. |