Quick BI AIPro trial guide

3.1 Prepare data

Select and upload one or two of your team's most frequently used core data tables. Examples include:

• Sales detail table (with fields such as date, region, category, sales amount, and profit)

• Ad campaign table (with fields such as channel, cost, impressions, clicks, and conversions)

• Order detail table (with fields such as order ID, customer, product, amount, and time)

Best practice: For best results, start with a single data table that contains complete fields, including a date dimension and other categorical dimensions.

3.2 Upload and enable the data experience skill

Important: The 'data experience' skill, which generates recommended test questions, must be manually uploaded and enabled in your Quick BI AIPro environment to use it. After preparing your data, download the skill package, upload it, and then proceed to the next step.

• Create a skill plan: My Skills

• Skill package: data-quickstart-v2.zip

3.3 Get recommended test questions

After you enable the data experience skill and upload your data, Quick BI AIPro automatically analyzes your data structure. It then generates 10 recommended test questions.

3.4 Run your first query

Start with questions from levels L1 and L2. Ask Quick BI AIPro questions in natural language to verify the accuracy of its answers.

Example questions:

• What were the total sales for each quarter in 2024?

• Which product category has the highest profit margin?

Validation criteria: Check if the data specification, units, and time range in the AI's answer match your expectations.

Phase objectives

• Core data tables prepared

• Data experience skill uploaded and enabled

• 10 recommended test questions generated

4.1 Involve business users

Invite one or two colleagues from each line of business—ideally those most familiar with the daily data—to act as pilot users. Examples include:

• Sales operations: Analyzes regional performance and creates daily/weekly reports.

• Marketing: Analyzes campaign effectiveness and channel ROI.

• Supply chain/Merchandising: Analyzes inventory, turnover, and categories.

4.2 Integrate into real workflows

Replicate your team's real-world tasks on Quick BI AIPro, such as querying data, creating tables, and writing reports. For example:

• Preparing data for the weekly sales report every Monday morning.

• Reviewing channel ROI at the end of each month.

• Sales attribution by category after a major promotion

Recommended practices:

1. Have business colleagues try asking questions in natural language instead of retrieving data manually.

2. Compare the answers from AIPro with the results from the previous method to verify accuracy.

3. Document which scenarios work reliably and which issues require optimization.

4.3 Sample questions

Based on the data from phase one, consider the following business questions:

• "What is the sales target achievement rate for each region this week?"

• "What is the ROI ranking for each channel last month?"

• "Which SKUs have inventory turnover days of more than 60 days?"

• "What are the reasons for the year-over-year decrease in sales for a specific region?"

Goals for this phase

• Involve at least 2 business units in the trial.

• Validate 3 to 5 real-world business scenarios.

• Create an "AIPro Applicable Scenarios List".

5.1 Focus on high-frequency scenarios

Review the usage records from the past two weeks. Find the 3 to 5 most frequent analysis questions and verify the following:

• Whether complex questions (such as those involving multi-dimension analysis, year-over-year, and period-over-period comparisons) are answered consistently.

• Whether the attribution for anomaly-related questions (such as "Why did a specific metric decrease?") is reliable.

• The effectiveness of report generation questions (such as "Generate a sales trend analysis report").

5.2 Consolidate reusable conclusions

Organize each successfully verified scenario into a three-part record: Question → AI Answer → Business Conclusion. For example:

• Question: "Why did sales in the East China region decrease after Q3 2024?"

• AI Answer: "The decrease was mainly due to the home appliances category. In Q4, sales for this category in the East China region dropped from 148,000 to 32,000."

• Business Conclusion: "Focus on the seasonal fluctuations and inventory strategy for the home appliances category in the East China region."

5.3 Assess the value

Assess the value of AIPro for your team based on the following dimensions:

• Efficiency: Whether tasks that used to take 30 minutes, such as data retrieval and table creation, are now reduced to 1 to 2 minutes.

• Coverage: How many departments and individuals have started using the tool.

• Ease of use: Whether business team members can ask questions independently without learning SQL.

Goals for this phase

• Identify 3 to 5 high-frequency, reusable scenarios.

• Establish a standard, reusable query format for each high-frequency scenario.

• Perform a preliminary assessment of the efficiency improvements provided by AIPro.

6.1 Trial review

Review the trial period with your team:

• Which scenarios were successful? Which ones require data or configuration optimization?

• What is the feedback from business colleagues?

• Is the accuracy of AIPro's answers sufficient for daily use?

6.2 Decide whether to subscribe

Based on the trial results, decide whether to continue using AIPro:

• If the trial demonstrates clear value, proceed with the subscription process.

• If some scenarios are not yet successful, contact the Quick BI team to extend the trial period or request targeted optimizations.

6.3 Future optimizations

After subscribing, make ongoing optimizations:

• Add more data tables to enable cross-table analysis.

• Build a library of standard questions to create shared AI-powered Q&A templates for your team.

• Create dedicated dashboards for specific business cycles, such as major promotions or monthly reports.

Goals for this phase

• Complete the trial review.

• Decide whether to subscribe.

• Identify the initial key use cases after subscribing.