Skip to main content
You are viewing help content for pre-release software. This document and the features it describes are subject to change.
All docs
V25.2
  • AI-powered Report Generation in Visual Studio Designer

    • 3 minutes to read

    AI-powered Report Generation allows you to create report layouts directly within the Visual Studio Report Designer. With the DevExpress AI-powered Assistant, you can connect to Azure OpenAI or OpenAI services and automatically generate reports based on natural language prompts.

    DevExpress AI Assistant for Visual Studio

    The DevExpress Unified Component Installer installs the AI-powered Assistant extension for Visual Studio.

    DevExpress AI-powered Assistant for Visual Studio

    Configure AI Services

    Go to ToolsOptionsDevExpressAI Assistant to open AI Assistant options.

    DevExpress AI Assistant Options

    Option Description
    Endpoint URL The service endpoint used to connect to Azure OpenAI. For OpenAI connections, this option must be empty.
    API Key A personal access key required to authenticate requests to the AI service.
    Model Name The identifier of the AI model (for example, gpt-4o-mini).
    Enable Logging If enabled, the extension records request and response details for diagnostic purposes. Logs are stored in the following folder: C:\Documents and Settings<UserName>\AppData\Local\DevExpress\AI_Assistant_VSIX
    Enable Prompt Caching If enabled, previously used prompts and responses are cached to improve performance and reduce repeated requests.

    Warning

    Some OpenAI models (for example, GPT-5, o1, and o3 series) do not support the Temperature parameter. To avoid runtime exceptions when targeting these models, set the Temperature to 1. The value is ignored by the service, but this setting prevents validation/compatibility errors.

    Configure Prompt to Report Options

    Option Description
    Fix Layout Errors When enabled, the AI automatically attempts to correct layout issues (for example, misaligned controls or overlapping elements) in the generated report.
    Retry Attempt Count Specifies how many times the AI should retry report generation if the initial attempt fails.
    Temperature Controls randomness in AI output and produces more deterministic results. Higher values increase variation. The default value is 0.

    Generate Reports with AI

    Add a DevExpress Report item or open the Report Wizard for an existing report using the report’s smart tag menu. In the Report Wizard, select AI Prompt-to-Report and click Next.

    AI Prompt-to-Report - DevExpress Report Wizard

    Select Data Source Option

    No Data Source
    Creates a complete report structure based only on the user’s natural language description.
    Add Data Source
    Creates and configures a report data source, and automatically includes this metadata in the LLM prompt.

    Configure Report Prompt

    Enter a prompt or choose a predefined prompt and modify it as needed. Then click Finish.

    Report Description - DevExpress Report Wizard

    Tip

    Output quality depends on the detail level in your prompt. Specify layout preferences, calculations, grouping, and visualization types. As with other LLM-based systems, results may require adjustments.

    Use the Temperature setting to control output randomness and RetryAttemptCount to specify how many times the system tries to fix layout errors in the LLM response.

    Troubleshooting

    Report Wizard Does Not Display AI Prompt-To-Report Item

    Possible Cause
    The DevExpress AI-powered Assistant for Visual Studio extension is disabled.
    Solution
    Open the Extension Manager in Visual Studio (ExtensionsManage Extensions…). Switch to the Installed tab and enable the DevExpress AI-powered Assistant for Visual Studio extension. Restart Visual Studio.

    HTTP 400 (Unsupported Value)

    Possible Cause
    Some OpenAI models (for example, GPT-5, o1, and o3 series) do not support the Temperature parameter.
    Solution
    Open AI Assistant Options and check Model Name and Temperature settings. If the selected model does not accept the Temperature, set the Temperature to 1 to avoid errors, or switch to a supported model (for example, gpt-4o-mini).

    HTTP 404 (Deployment Not Found)

    Possible Cause
    The specified model deployment does not exist. This may happen if the deployment name is incorrect, if the deployment was recently created and is not yet active, or if it was deleted.
    Solution
    Open AI Assistant Options and check the Model Name setting.
    See Also