Quick BI supports the Q Chat feature. Using the intelligent data assistant Q Chat, combined with ChatBI's intelligent interaction capabilities, you can directly obtain data results through natural language interaction. This enables instant data Q Chat, empowering everyone to analyze data and driving new ways of data consumption. In the Q Chat conversation interface on PC or mobile, preview and select datasets, directly enter questions or quick questions in the input box, engage in multi-turn conversations, and view historical Q Chat conversations in the conversation list.
This topic describes how to use the Q Chat feature.
Prerequisites
You have created a dataset and defined the query configuration.
You have Ask Data resource permissions. For specific steps, see Permission Management.
PC Access Points
Access Point 1: On the Quick BI home page, click Intelligent Q Chat, and then select Q Chat in the Intelligent Q Chat super box.
Access Point 2: In the navigation pane on the left of the Intelligent Q Chat home page, select Q Chat.
The two access points offer similar functions. This topic uses Access Point 2 as an example.
Data Sources
Datasets
All datasets and datasets under each analysis subject for which you have Q Chat permissions.
For information about configuring Qwen dataset permissions, see Qwen Permission Management.
For information about configuring analysis subjects, see Analysis Subject Management.
Data with the "Sample" flag in the upper-left corner are sample datasets.
If you do not need sample data, you can disable it in Whether to display sample datasets.
Data Files
Uploaded local data files.
Upload Data Files
On the Data Files tab, if no files have been uploaded, click Upload File to upload them.
If data files have already been uploaded, click the Upload File icon in the upper-right corner to access the upload interface.
If more than five data files are uploaded, click All Data Files to view them. Then, on the Data Files tab, click Upload File to access the upload interface.
On the file upload interface, click or drag files to this area to upload them.
Preview data.
On the data preview interface, set the Display Name for the dataset.
NoteEnter a standard name. Do not use special characters or leading/trailing spaces.
Supports data preview and viewing field details.
After confirming, save the data. Options include Save and Start Q Chat or Save Only.
Save and Start Q Chat: Selects the data after clicking.
Save Only: Does not select the data after clicking.
After a successful upload, select the data file to start a Q Chat session.
Manage Data Files
Configuration
Hover over the data file, click the
icon on the right, and select Configure.
If more than five data files are uploaded, click All Data Files. Then, on the Data Files tab, hover over the data file, click the icon on the right, and select Configure.
On the data file configuration interface, modify the Display Name, perform a data preview, and view field details.
Delete
Hover over the data file, click the
icon on the right, and select Delete.
If more than five data files are uploaded, click All Data Files. Then, on the Data Files tab, hover over the data file, click the
icon on the right, and select Delete.
In the confirmation dialog box, click OK to delete the uploaded data file.
Preview and Select Datasets
Select a Single Dataset
On the Q Chat conversation page, view all datasets and datasets under each analysis subject, along with uploaded data files. Hover over a target dataset to preview it or directly ask a question based on it.
Click Preview (①) to view the dataset's field details and data preview. You can also perform a quick question or click Ask a Question in the upper-right corner to enter the Q Chat dialog box.
Field Details
Data Preview
Hover over a dimension field to view its details.
Quick Question
Ask a Question
Perform data Q Chat based on the current dataset.
Click Ask a Question (②) to enter the data Q Chat conversation interface and perform data Q Chat based on the current dataset.
After selecting data, click the
icon to the right of the selected data to preview data fields and view key metrics and analysis dimensions.
Supports expanding or collapsing data fields.
During data Q Chat, data preview and dataset switching are supported.
Preview Data
Click the
icon to the right of the selected data to preview data.
Dataset Switching
Click the name of the selected data to view available data.
Hover over the target dataset and click Ask a Question to switch datasets.
Alternatively, preview the target dataset first, then ask a question.
Select Multiple Datasets
When your question matches multiple datasets, select multiple datasets for Q Chat.
This section describes only the procedure for selecting multiple datasets. Other operations are the same as selecting a single dataset.
Enter a question.
The system displays data potentially related to the question. Select the required data from the options.
Check multiple datasets and click Ask a Question Based on Selected Data.
Obtain analysis results.
Ask Questions and Get Answers
XiaoQ Wenshu supports multiple inference and data interpretation methods. This example uses the default official inference and data interpretation method for demonstration. For analysis results of other methods, see the following Inference and Interpretation Methods.
In the dialog box, enter and send your question. Options include direct question input, quick questions, and speech input.
Direct Question Input
For example, enter "What is the proportion of order quantities for each order level?" After sending, the system returns data analysis results and displays the analysis process on the right.
If you do not need to view the analysis process, click Analysis Process Expanded to hide it.
When hidden, the results only display the chart.
Click Analysis Complete to display the analysis process again.
When you enter content, recommended fields and recommended questions may appear. For example, entering "2024" displays the following recommended fields and questions. Select as needed.
Question Clarification
After you enable this feature, when your question meets the conditions for follow-up question clarification—for example, missing time or metric elements, or ambiguous scenarios involving multiple metrics—Smart Q will ask a follow-up question to confirm the required elements before providing a more accurate answer.
In question clarification, switch the relevant period and analysis metrics, then click Start Analysis.
Alternatively, modify the question based on the plan.
Alternatively, select Skip and Answer Directly.
Favorite Questions
Supports favoriting up to 30 questions.
Questions asked using uploaded data files cannot be bookmarked.
Click the
icon to the left of a question to favorite the question.
After favoriting, use it in quick questions.
When no data is selected, click a favorited question directly to ask it.
Click the
icon again to remove from favorites.
Copy Question
Click the 
icon to the left of a question to copy the question.
Quick Questions
Click the
icon in the lower-left corner of the question box to enter the quick question interface. Then, click Recommended, Favorited, or Recently Used questions to ask them.For favorited questions, check Automatically switch datasets. This option is checked by default. When checked, clicking a favorited question switches to the bound dataset and asks the question. When unchecked, the question is asked using the current dataset.
Alternatively, click a recommended question above the question box to ask it.
Click the
icon to get a new set of recommended questions.
After each conversation, directly select one of the three recommended questions below the chart to ask it.
Speech Input
After you enable the speech Q Chat configuration, you can use speech input to ask questions.
Click the Speech Input icon to enter the speech conversation interface.
On the speech conversation interface, ask a question.
The system recognizes your speech and converts it to text.
Click the
icon on the left to cancel input, or click the 
icon on the right to switch to text input.Click Send or press Enter. The system returns data analysis results.
View Results
When viewing data, the following features are supported.
Data screening (①): The chart displays the query conditions for the current data retrieval. Supports switching.
Chart switching (②): Switch the visualization display to better suit viewing habits (the range of switchable options varies based on the data).
View AI data retrieval process (③): Verify if the data retrieval results meet your requirements. Supports viewing business logic SQL and execution SQL.
Fluctuation Attribution and Prediction Trend (④)
Fluctuation Attribution
NoteScorecards and line charts support fluctuation attribution. The fluctuation attribution feature appears only if the following conditions are met:
Supported chart types: scorecards and line charts that are not prediction line charts themselves.
The chart has only one dimension.
The dimension is of date type and has no legend.
The chart for which fluctuation attribution is required has no errors.
Line chart conditions: 1 date dimension, 1 to 3 measures.
Scorecard conditions: 1 date dimension, 1 measure.
Click the
icon in the upper-right corner to configure the attribution report.
You can modify attribution metrics, data interpretation approach, and model selection. After you click OK, the system automatically updates the attribution report. For related operations, see Metric Insights.
Add the attribution report to the Q Chat report as needed.
Prediction Trend
NoteLine charts and bar charts support the prediction trend feature, which appears only when the following conditions are met:
Supported chart types: Line charts and bar charts that contain a date and do not have a legend.
Line charts and column charts that are not prediction results themselves.
The chart to be predicted has no faults.
The number of dimension values in the chart to be predicted is greater than or equal to 12.
Line chart display conditions: one date dimension and 1 to 3 measures.
Supports clicking the
icon in the upper-right corner to share the prediction trend.
Full screen view (⑤).
Rename (⑥)
Supports renaming the returned results. Alternatively, click the red box directly to rename.
Share data (⑦)
Copy the link to analyze or share it with IM software.
NoteTo share in IM software, bind an enterprise organization and a personal account.
Supports sharing to WeCom, DingTalk, and Lark.
Export data (⑧)
Supports exporting to local in Excel file format. Supports customizing the export name.
NoteOnly PC supports export; mobile does not.
Data Interpretation (⑨)
NoteInitiate custom data interpretation here. Interpret Q Chat results according to your specified interpretation ideas and models.
Refer to the interpretation ideas in the example. Customize your interpretation ideas and select a model. Click Generate Interpretation, and the system automatically interprets the data for you.
Supports stopping interpretation during result return.
Supports adopting or modifying interpretation results. After adoption, the following data interpretation report is generated.
To modify data interpretation ideas or models for a second interpretation, click the
icon to the right of data interpretation to modify and generate a new data interpretation report.
Like/Dislike (⑪)
Click
to like, click 
to dislike. When disliking, provide feedback.
More (⑫)
Copy Trace_ID
Add to Smart Q Report
Clear Related Memory
If the answer is unsatisfactory, click Clear Related Memory to clear the memory of the current question and similar content, then ask again to get a new answer.
NoteOnly file queries support clearing relevant memory.
Inference and Interpretation Methods
Method Description
Inference and interpretation methods default to the system's built-in models. Supports selecting system built-in Large Language Models (LLMs) and custom models. Select appropriate inference and interpretation methods to analyze your data as needed.
Examples
Using the question: "What is the proportion of order quantities for each order level?" as an example, here are examples of different inference and interpretation methods:
Example 1: Select the default official inference and no data interpretation.
The system returns the analysis process and chart.
Example 2: Select the built-in Qwen-Max model for both inference and data interpretation.
The system uses the Qwen-Max model to return the inference process, data interpretation results, and chart.
Example 3: Select the built-in DeepSeek-R1 model for both inference and data interpretation.
The system uses the DeepSeek-R1 model to return the inference process, data interpretation results, and chart.
Follow-up Questions
Use the follow-up question feature to continue asking questions about the results of the previous question, gaining a more precise understanding of the data. Here's an example:
A question supports only one follow-up question.
On the conversation interface, click Follow-up Question to ask a follow-up question about the previous question: "Which province has the highest sales?"
Enter the follow-up question as needed. For example: What is the best-selling product?
Click the
icon in the lower-right corner of the dialog box or press Enter on the keyboard to send the follow-up question and obtain analysis results.
NoteThe system automatically exits follow-up question mode. To continue asking follow-up questions, enable the follow-up question button again.
Conversation List
Click the
icon in the upper-left corner to view historical Q Chat conversations.
Click the
icon to enter a new Q Chat interface.
Display on Mobile Terminal
Mobile Micro-App Configuration Entry
Follow the steps in the diagram to enter the mobile editing interface.
Configure Smart Q Data Query for mobile micro-apps in the editing interface.
Suggested Question Formats
When asking questions, add time limits. For example:
"Proportion of sales amount for each product" should be rewritten as "Proportion of sales amount for each product this year"
In scenarios involving multiple dimension value aggregations, ambiguity can arise between individual and total calculations. If separate calculations are needed, add keywords such as "respectively," "per X," "each X," "total sum," or "total" for distinction. For example:
"Sales in Zhejiang and Jiangsu", "Sales from 2020 to 2023" currently need to be rewritten as "Sales in Zhejiang and Jiangsu respectively", "Sales per year from 2020 to 2023", "Sales for each year from 2020 to 2023", "Total sales for Zhejiang and Jiangsu", "Total sales for Zhejiang and Jiangsu"
In scenarios involving maximum/minimum values for multiple dimension values, ambiguity can arise between individual and total calculations, or whether the maximum/minimum refers to grouped or ungrouped data. Rephrase these into: maximum/minimum for a single dimension value, or window functions for dimension maximum/minimum. For example:
"Which product sold best in Zhejiang and Jiangsu?" currently needs to be rewritten as "The product with the highest sales in Zhejiang", "The product with the highest sales in Jiangsu", "The product with the highest sales in each province"
"The 3 best-selling months last year in Zhejiang and Jiangsu" currently needs to be rewritten as "The 3 best-selling months last year in Zhejiang", "The 3 best-selling months last year in Jiangsu", "The top 3 order months by sales in each province"
For scenarios with ambiguous combinations of dimension values, explicitly specify the dimension when entering the question. For example:
"Products purchased by corporate clients, ranked by sales". If "corporate" (dimension value) exists within "client type" (dimension), "corporate clients" might be ambiguously interpreted as "our company's clients" due to other semantic interference. Therefore, explicitly specify: "Products purchased by clients of type 'corporate', ranked by sales"