Creating custom commands
Turn a reusable prompt into a custom command that lives inside an assistant, then run it on demand — with scheduled runs coming soon.
Draft — pending review. This article was drafted from JoySuite's product marketing and has not yet been verified against the shipping product. Feature names, screens, and steps may differ.
Most teams have a "perfect prompt" that lives in one person's head — the exact wording that produces a policy summary, a territory brief, or a deal review the way your organization likes it. Custom commands turn that prompt into a reusable command that lives inside an assistant, so everyone runs it the same way and gets the same quality every time.
Unlike the built-in commands that ship with each of the default assistants, a custom command is one you build from your own prompt and wire to your own knowledge sources. Once it lives inside an assistant, there is nothing to copy, paste, or search for — it is always there, ready to run.
Try running a command here:
How custom commands are shared
Where a command lives determines who can use it.
- Team commands are built for a specific assistant and shared with everyone who has access to that assistant. They inherit the assistant's access — so there are no links to share and no "where was that posted?" A team command never changes the assistant's sources or guardrails; it is a capability layered on top of the governed foundation.
- Personal commands are shortcuts you create for yourself. They are scoped to your account and invisible to everyone else — your own prompts, always at hand, with no admin required.
Because a team command inherits the assistant's access, updating it once updates it for everyone. There is no version drift and nobody is left running an old copy of the prompt.
Create a custom command
The exact wording of buttons and menus may differ in your workspace, but the flow is the same.
Open the assistant's settings
Choose the assistant where the command should live, then open its command or settings area. Personal commands can be scoped to just you; team commands are shared with everyone who has access to that assistant.
Start a new command and name it
Give the command a short, memorable name — this is what people will type after the forward slash, so keep it clear (for example, a "territory brief" or "deal risk scan").
Write the prompt template
Paste in your best prompt. Where the command should ask for a detail each time it runs, add an input the runner fills in — such as the territory name or the document to review.
Connect knowledge sources
Point the command at the files, FlexDocs, or connected systems it should draw on, so answers are grounded in your content rather than general knowledge.
Save it to the assistant
Save the command. It now appears in the slash-command list for everyone with access to that assistant (or just for you, if it is personal).
Running a command
Once a command exists, anyone with access can run it on demand — they type the command in a chat and get the output immediately. Reps run a territory brief before a call; new hires run a benefit-options command during enrollment. See Using a command for how running one works in a chat.
Coming soon. Scheduled command runs are planned but not yet available. The goal is to let a command run automatically on a schedule — for example, a weekly pipeline summary — and deliver the result to a channel or inbox. Until then, commands run on demand.