UpKeep Nova offers three different modes so you can explore, test, and run tasks with the right level of safety and control. Choosing the right mode helps you balance learning, accuracy, and speed.
Playground Mode — Safe Exploration
Playground Mode is the best way to understand how the bot works. Because it runs only on simulated data, nothing in your account will change.
What it does
Runs in a sandbox environment with simulated data
Lets you see how the agent interprets prompts and structures actions
No changes are saved to your real account
When to use it
Exploring how the agent thinks
Learning to prompt without risk
Training new team members or demoing
Example prompts
“Simulate creating a new warehouse location with 10 assets”
“Show me how you would triage all pending requests”
“Preview a preventive maintenance schedule for my top 3 assets”
Ask Mode — Guardrails with Control
Ask Mode is where you start using the bot in your real account while keeping full control over every step. You decide what gets confirmed and saved.
What it does
Prepares a plan for your request
Shows each action in the sidebar for you to confirm step by step
No action is taken until you approve it
When to use it
Running real actions with extra control
Reviewing and approving every change before it’s saved
Building trust in the agent with your data
Example prompts
“Create a new facility and add 5 assets — ask me to confirm each step”
“Triage my pending requests but show me the plan first”
“Propose PMs for my top 3 assets and let me approve before saving”
Agent Mode — Full Autonomy
Agent Mode is where you let the bot run at full speed. Once you send a prompt, the agent executes immediately without step-by-step confirmation.
What it does
Executes your requests immediately after you send them
Applies changes directly to your account
⚠️ Important: Deletions (e.g., deleting a work order, request, or asset) always require your approval, even in Agent Mode
When to use it
Confident users who know exactly what needs to be done
Repeatable tasks the agent has proven reliable with
High-volume actions where speed matters most
Example prompts
“Triage my pending requests”
“Create PMs for my top 5 assets this quarter”
“Decline all duplicate requests from this list”
Quick Comparison
Mode | Safety Level | Best For | When to Use |
Playground | 100% safe (no real changes) | Learning & testing | Use when you’re exploring how the bot works and want to understand its capabilities without risk |
Ask | High (you confirm each step) | Early production use | Use when you want to see exactly what will happen before applying changes |
Agent | Lower (no confirmations, except deletions) | Confident, repeat users | Use when you know exactly what needs to be done and want it applied immediately — deletions still always require approval |
✨ By choosing the right mode, you can safely learn, stay in control, or move fast with autonomy. Start in Playground, step up to Ask for real control, and switch to Agent once you’re confident. Remember: deletions always require confirmation in every mode.