Configure AI functions in Quick BI spreadsheets-Quick BI(Quick BI)-阿里云帮助中心

Quick BI provides AI functions in spreadsheets, allowing spreadsheet creators to directly use AI features such as translation, classification and labeling, information extraction, and summarization. You can intelligently process text data without writing complex formulas. This significantly lowers the barrier to data processing and improves analysis efficiency. This topic describes how to configure AI functions in a spreadsheet.

Notes

You can use AI functions only after an organization administrator enables the feature on the organization management page. Otherwise, the AI function option is grayed out and inaccessible. For more information, see AI function management.
AI functions use the default large model configured in Organization Management. To change the model, contact your organization administrator. For detailed instructions, see Large Model Configuration.

Entry points

You can access AI functions from the following entry points on the spreadsheet editing page.

Toolbar entry: Click the AI functions icon () on the toolbar and select an AI function from the drop-down list.
Cell entry: Right-click a cell, hover over Create AI function, and then select an AI function from the submenu that appears.

Procedure

Select an AI function from the list and configure its rules. You can use the classification and labeling, information extraction, translation, and summarization functions. If the standard functions do not meet your analysis needs, you can configure a custom AI function. The following sections describe each function and provide configuration instructions.

Classification and labeling

The classification and labeling function automatically identifies text content and applies predefined category labels. This function is useful for quickly standardizing and categorizing text data.

For example, when analyzing customer satisfaction, you can use the classification and labeling function to automatically categorize customer feedback. This helps you assess customer satisfaction more efficiently and intuitively. This section uses this example to explain how to configure the classification and labeling function.

In the Customer Feedback Information table, select the text area that you want to process. Click the icon on the toolbar and select Classification and labeling.

In the Classification and labeling dialog box, configure the following parameters.

Parameter	Description
Select data range	Select the data range to classify and label. The currently selected cell range is filled by default. To modify the range, click the icon and select a range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note The selected range must be in a single column.
Category options	Set the label categories for classification. Option 1 is provided by default, and you can edit its name as needed. You can also perform the following actions: ① Click Add to create a new label option. ② To add or manage multiple category labels at once, click Batch Edit to quickly import, delete, or modify them. ③ To remove a label option, click the Delete icon.
Custom classification rules	Enter semantic rules for each category label to help the AI classify text data more accurately. Describe the meaning of each category label. You can provide typical keywords or examples for the AI to learn from. For this example: Satisfied: The customer expresses approval or appreciation for the product. Neutral: The customer's overall sentiment is not strong; they offer suggestions for improvement but do not express dissatisfaction. Dissatisfied: The customer complains about the product, service attitude, or other issues.
Preview	You can preview the labeling result for the first data entry under the current rules to quickly validate the rules. Click Preview. The system runs the classification and labeling function on the first data entry in the selected range and returns the result to the preview box.
Select result location	Set the output location for the classification and labeling results in the spreadsheet. The following options are available. To the right of the referenced column: Inserts the AI results to the right of the referenced data column. Custom: Manually select the location for the AI results. Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note Ensure that the selected custom area does not contain any data. The size of the output range must match the size of the input range.
Auto-update	Specify whether the classification and labeling results automatically update when the source data changes. This feature is disabled by default. In this case, the output result does not change when the content of the referenced cell changes. If you enable this feature, the system reruns the AI function to update the result when the content of the referenced cell changes.

After you complete the configuration, click OK. Once the AI processing is complete, you can view the classification and labeling results in the spreadsheet.

Information extraction

The information extraction function extracts specific information from text.

For example, when you export sales data from a business system for analysis, key information such as customer names and order numbers might be mixed within a single block of text. In this case, you need to intelligently extract these key fields from the original order information to display the sales data in a structured and clearer way. This section uses this example to explain how to configure the information extraction function.

In the Order Information table, select the text area you want to extract information from. Click the icon on the toolbar and select Information extraction.

In the Information extraction dialog box, configure the following parameters.

Parameter	Description
Select data range	Select the data range for information extraction. The currently selected cell range is filled by default. To modify the range, click the icon and select a range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note The selected range must be in a single column.
Content to extract	Describe the specific content you want the AI to extract from the text data. For example: "Extract the order number."
Custom reference examples	Provide pairs of original text and expected output to improve the AI's extraction accuracy. We recommend that you follow the format provided by the system. For example: From the text "Order No.: ORD20240501", extract the order number "ORD20240501".
Preview	You can preview the extraction result for the first data entry under the current rules to quickly validate the configuration. Click Preview. The system runs the information extraction function on the first data entry in the selected range and returns the result to the preview box.
Select result location	Set the output location for the extracted results in the spreadsheet. The following options are available. To the right of the referenced column: Inserts the AI results to the right of the referenced data column. Custom: Manually select the location for the AI results. Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note Ensure that the selected custom area does not contain any data. The size of the output range must match the size of the input range.
Auto-update	Specify whether the information extraction results automatically update when the source data changes. This feature is disabled by default. In this case, the output result does not change when the content of the referenced cell changes. If you enable this feature, the system reruns the AI function to update the result when the content of the referenced cell changes.

After you complete the configuration, click OK. Similarly, extract the customer name and order amount information. The final result is shown below.

Translation

The translation function translates text between languages. It is useful for processing multilingual information and for cross-language team collaboration.

For example, after collecting feedback from overseas users, you can use the translation function to automatically translate English content into Chinese, allowing the local business team to efficiently understand user opinions. This section uses this example to explain how to configure the translation function.

In the Overseas User Feedback table, select the text area that you want to translate. Click the icon on the toolbar and select Translation.

In the Translation dialog box, configure the following parameters.

Parameter	Description
Select data range	Select the data range for translation. The currently selected cell range is filled by default. To modify the range, click the icon and select a range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note The selected range must be in a single column.
Target language	Select the target language for the translation. Supported languages include English, Simplified Chinese, Traditional Chinese, and Japanese.
Custom translation requirements	Enter custom requirements for translation style, phrasing, and text length to enhance the professionalism and suitability of the results. For example: "Translate in a formal style, with a maximum length of 30 characters."
Preview	You can preview the translation result for the first data entry under the current rules to quickly validate the configuration. Click Preview. The system runs the translation function on the first data entry in the selected range and returns the result to the preview box.
Select result location	Set the output location for the translation results in the spreadsheet. The following options are available. To the right of the referenced column: Inserts the AI results to the right of the referenced data column. Custom: Manually select the location for the AI results. Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note Ensure that the selected custom area does not contain any data. The size of the output range must match the size of the input range.
Auto-update	Specify whether the translation results automatically update when the source data changes. This feature is disabled by default. In this case, the output result does not change when the content of the referenced cell changes. If you enable this feature, the system reruns the AI function to update the result when the content of the referenced cell changes.

After you complete the configuration, click OK. Once the AI processing is complete, you can view the translation results in the spreadsheet.

Summarization

The summarization function intelligently creates abstracts and extracts key points from long texts. It is useful for scenarios where you need to quickly understand the main ideas of large amounts of text and improve information retrieval efficiency.

For example, when analyzing work order data, you can use the summarization function on customer feedback to summarize their requests, helping the business team quickly grasp core user needs and pain points. This section uses this example to explain how to configure the summarization function.

In the Order Information table, select the text area that you want to summarize. Click the icon on the toolbar and select Summarization.

In the Summarization dialog box, configure the following parameters.

Parameter	Description
Select data range	Select the data range to summarize. The currently selected cell range is filled by default. To modify the range, click the icon and select a range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note The selected range must be in a single column.
Custom summary requirements	Enter custom requirements for the summary's style, length, and key focus areas to improve its usefulness and semantic consistency with the original text. For example: "Limit the summary to 10 words, ensuring it is concise and easy to understand."
Preview	You can preview the summary result for the first data entry under the current rules to quickly validate the configuration. Click Preview. The system runs the summarization function on the first data entry in the selected range and returns the result to the preview box.
Select result location	Set the output location for the summary results in the spreadsheet. The following options are available. To the right of the referenced column: Inserts the AI results to the right of the referenced data column. Custom: Manually select the location for the AI results. Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note Ensure that the selected custom area does not contain any data. The size of the output range must match the size of the input range.
Auto-update	Specify whether the summary results automatically update when the source data changes. This feature is disabled by default. In this case, the output result does not change when the content of the referenced cell changes. If you enable this feature, the system reruns the AI function to update the result when the content of the referenced cell changes.

After you complete the configuration, click OK. Once the AI processing is complete, you can view the summarization results in the spreadsheet.

Custom AI function

If Quick BI's standard AI functions do not meet your needs, you can use natural language to call a large model and create custom analysis tasks.

For example, after receiving after-sales reviews, the support team needs to analyze the review content and rating to quickly identify which feedback items are high-priority, which are logistics-related issues that can be forwarded to the shipping carrier, and which are routine feedback that can be handled later. In this case, you can use a custom AI function to set comprehensive rules and automatically determine the processing priority for each review. This section uses this example to explain how to configure a custom AI function.

In the Customer Information table, select the text area to use as function variables. Click the icon on the toolbar and select Custom.

In the Custom AI function dialog box, configure the following parameters.

Parameter	Description
Custom variables	Define data variables that can be referenced by the AI task. By default, the selected data is added as Variable 1. You can change its name as needed. You can perform the following actions: ① Click New Variable to add another variable. ② Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note The selected range must be in a single column. ③ To remove a variable that is no longer needed, click the Delete icon. If this variable is referenced in the AI task description, the reference will become invalid if the variable is deleted.
AI task description	Describe the AI task in natural language. While writing, you can click Reference Variable to include variables you have configured under Custom variables in the task description.
Preview	You can preview the output for the first data entry under the current rules to quickly validate the configuration. Click Preview. The system runs the custom function on the first data entry in the selected range and returns the result to the preview box.
Select result location	Set the output location for the results in the spreadsheet. The following options are available. To the right of the referenced column: Inserts the task results to the right of the rightmost variable column in the spreadsheet. In this example, the results will be inserted next to the Partner business column. Custom: Manually select the location for the AI results. Click the icon and select a data range by dragging the mouse in the spreadsheet, or manually edit the formula in the Cell range input box. Note Ensure that the selected custom area does not contain any data. The size of the output range must match the size of the input range.
Auto-update	Specify whether the output results automatically update when the source data changes. This feature is disabled by default. In this case, the output result does not change when the content of the referenced cell changes. If you enable this feature, the system reruns the AI function to update the result when the content of the referenced cell changes.

After you complete the configuration, click OK. Once the AI processing is complete, you can view the execution results of the custom function in the spreadsheet.

Function management

After adding an AI function, you can manage it in the following ways on the spreadsheet editing page:

① Regenerate: Click the icon to call the AI function and regenerate the entire result area.
② Edit: Click the icon to open the edit dialog box to adjust the function's data range, calculation rules, and other parameters.
③ Delete: Click the icon and confirm in the dialog box to delete the entire column of AI function results.
④ Update selected cells: If you are not satisfied with an individual cell's result, select the cell and click the icon on the right to regenerate it without updating the entire area.

If an AI function fails to generate a result, you can regenerate the failed area.

Click the icon to regenerate all failed cells in the current area.
Hover over a failed cell and click Retry this cell in the error message to regenerate only that cell. You can also click Retry failed items to regenerate all failed cells in the result area.