Testing your setup in the AI Playground

The AI Playground in CloudRadial AutomationAI is a test surface for your installed extensions and runners. You give an agent a goal, and it works toward it using your own AI provider, your installed extensions and the secrets in your runner's Key Vault, then reports back. It tells you whether your tools can actually do the job before you build a workflow. Choose which extensions the agent may use, watch every tool call as it happens, and convert a run that works into a workflow. This article explains how to start a run, scope the tools it may use, follow its live activity, and cancel a run in progress. The AI Playground is available to all roles.

The AI Playground is not a setup guide or a general-purpose assistant. It cannot install extensions or add secrets for you, and a goal that names a tool you have not installed and configured will fail.

The AI Playground was previously called Ask AI. Existing links to /ask-ai still work and redirect to the new location.

In this article:

• What the AI Playground Is
• Before You Start
• Starting a Run
• Scoping the Runner and Tools
• Following the Activity Timeline
• Cancelling a Run

What the AI Playground Is

The AI Playground runs a single, goal-directed agent task. You describe a goal in plain language, and the agent reasons about it, calls the tools you have made available, and returns a result. Each run is one-shot rather than an ongoing chat. While a run is in flight the composer is hidden, making it clear that the agent is working toward the goal you gave it, not queuing follow-up messages.

Open it from AI Playground (/ai-playground) in the sidebar.

Before You Start

The AI Playground runs against what you have already set up, so complete these first:

• A registered runner that is reporting healthy. See Deploying a runner from the Azure Marketplace
• Access to your runner's Key Vault. See Accessing and updating your runner's Key Vault
• An AI provider set on the runner. See Choosing the runner AI provider
• The extensions your goal needs, installed with their secrets in the Key Vault. See Browsing and installing extensions

Starting a Run

The launcher opens with a goal composer.

1. Type a goal into the composer, for example "Show all users in Microsoft 365 in CSV format".
2. Pick a runner and select the extensions whose tools the agent may use.
3. Select Send (or press Cmd/Ctrl+Enter) to start the run.

Quick-start suggestions appear below the composer; selecting one fills the goal box. Your recent goals can be recalled in the composer with the up and down arrow keys.

Scoping the Runner and Tools

Before sending, set the run's scope with the two fields above the composer:

• Runner: the runner this run executes on. If a runner is offline, the run waits until it next polls. Pick another runner to run immediately.
• Tools: the extensions whose tools the agent may use. Only the selected extensions' tools are sent to the model. An empty selection sends no extension tools at all, so you must select at least one extension before you can send.

The selection shows an approximate count of available tools. If the selected extensions expose more tools than the model's limit, a message asks you to narrow the selection before the run can dispatch. Your runner and extension choices are remembered for your next visit. If your workspace has companies registered, a Company field also appears, scoping the run to that company’s credentials; it is hidden when no companies are registered, because company scope has nothing to select.

Following the Activity Timeline

On send, the page switches to the active view. It shows the goal, a live Activity timeline, and the final result.

• Reasoning steps appear as quiet "Thinking" markers.
• Tool calls appear as named tool cards. A preview (dry-run) call is marked Preview, not executed.
• The final answer appears in the Result panel when the run succeeds.

If the run fails, times out, or is cancelled, a status entry appears with a link to the full audit. If the agent needs a decision before it can continue, the run parks and a link to Action Required appears. A live indicator shows whether the activity feed is connected; the run is always saved, so a dropped connection never loses the result. Select Full audit to open the run's complete detail, or New Prompt to return to the launcher for another goal.

Cancelling a Run

While a run is in progress, select Stop in the goal card to cancel it. The run is finalized as cancelled and any later result from the runner is discarded. Cancelled runs remain in your history with a link to their audit.

If you are still having trouble, we're here to help! Submit a ticket for assistance, and don't forget to check our status page to ensure there are no outages in your area.