Set up and use the ArcGIS Pro assistant (beta)
To set up and use the assistant, complete the following steps:
Ensure that ArcGIS Pro 3.8 is installed, including the Semantic Search feature listed in the AI Models section on the ArcGIS Pro Setup installation dialog box.
If ArcGIS Pro 3.8 is already installed, you can modify the setup to add these application features if necessary.
Install the ArcGIS Pro assistant 3.8 (beta) by downloading and running the setup provided on the ArcGIS Pro assistant Early Access Community site.
Once installation is complete, start ArcGIS Pro, and sign in to an ArcGIS Online organization that allows the use of assistants or an ArcGIS Enterprise organization that allows the use of assistants.
ArcGIS Online organizations must also allow the use of beta features.
Ensure that your role includes privileges to use AI assistants in ArcGIS Online or ArcGIS Enterprise as applicable.
In an open ArcGIS Pro project, on the ribbon, on the Help tab, in the Assistant group, click Assistant (Beta) to open the Assistant (Beta) pane.

Click OK to dismiss the Logging Notice if necessary.
By default, the assistant opens in ArcGIS Pro Help mode.
Optionally, choose a different mode for working with the assistant.
The following table identifies the available modes and how to access them:
Mode How to access ArcGIS Pro Help Enter a question in the window at the bottom of the pane to begin a help conversation. For more information about using the assistant in ArcGIS Pro Help mode, see Ask the ArcGIS Pro assistant (beta) for help. ArcGIS Pro Actions Click the ArcGIS Pro Actions card to perform an action using the assistant. For more information about using the assistant in ArcGIS Pro Actions mode, see Perform an action with the ArcGIS Pro assistant (beta). ArcPy Code Click the ArcPy Code card to enter prompts to create Python code snippets that use the ArcPy Python site package. For more information about using the assistant in ArcPy Code mode, see Generate Python and ArcPy code with the ArcGIS Pro assistant (beta). Graph Query Click the Graph Query card to select a knowledge graph and ask natural language questions about it. For more information about using the assistant in Graph Query mode, see Create a graph query with the ArcGIS Pro assistant (beta). Query Layer Click the Query Layer card to select database tables and ask natural language questions about them. For more information about using the assistant in Query Layer mode, see Create a query layer with the ArcGIS Pro assistant (beta). Note:
The assistant also has an Error Assistant mode that you can access only by clicking Ask Assistant
on certain error and warning messages. For more information, see Get help with error and warning messages from the ArcGIS Pro assistant (beta).Type a natural-language prompt in the prompt text box at the bottom of the pane.
Tip:
Alternatively, you can use voice input to speak your prompt. See the Use voice input section below.
Press Enter or click Ask the question
to submit the prompt.The conversation with the assistant begins, and the response displays in the pane. Before the response displays, you may see processing messages, such as tasks the assistant performs or datasets it accesses to fulfill the request.
Read the displayed response, or hover over it and click Read aloud to hear it read aloud.
You can pause or stop playback as needed.
Optionally, continue the conversation by entering and submitting follow-up prompts to the responses you receive.
Use voice input
With voice input, you can enter a question or a request as speech instead of typing it. The assistant transcribes your speech as text in the prompt text box.
Voice input is available in all assistant modes.
To enter a prompt using speech, complete the following steps:
Click Voice Input
or press Ctrl+Shift+Spacebar to start voice input.Speak the prompt into your microphone.
Click Voice Input
or press Enter when finished.Your speech is transcribed in the prompt text box.
Review the transcribed text and edit it as needed.
Press Enter or click Ask the question
to submit the prompt.
Customize audio settings
To customize audio settings for voice input and output, complete the following steps:
Click the menu button
, and click Assistant Settings
.Click the Properties tab, and click the Audio drop-down arrow to expand the section.
Optionally, adjust any of the following settings:
Setting How to adjust Allow Voice Input Check or uncheck the check box to globally enable or disable voice input. Submit prompt after a silent pause Check the check box to automatically submit your prompt after you stop speaking for the selected duration. Choose a longer pause duration if you need more time between phrases. Read out responses Check the check box to enable the assistant to automatically read new assistant responses aloud. Enable voice flow Check the check box to allow for a continuous conversation without needing to enable voice input each turn. After you speak and submit the first prompt, the assistant completes its response, and it immediately listens again for voice input. If Read out responses is enabled, it waits until playback finishes before listening again. The Allow Voice Input and Submit prompt after a silent pause settings are automatically enabled when you enable voice flow. Playback Speed Click the button of the desired playback speed for reading responses. Playback speeds range from 0.5x through 2x, in increments of 0.25x. Voice Click the drop-down arrow, and click any option to set the voice used when the assistant reads responses aloud.