Optimizing data for AI
IBM watsonx BI is a powerful analytics tool that uses large language models (LLM) to run complex, multi-step BI queries on your data. While watsonx BI has contextual understanding of data and can quickly respond to your questions, the quality of data determines the quality of the LLM’s generated responses.
To enhance the quality of the generated query and the final answer, it is important to provide clear and comprehensive context to the LLM.
The following factors are used by the LLM to understand questions and the context of your data when generating query statements:
- Asset name
- Asset description
- Column name
- Column description
- Column identifier
- Column data type, usage, aggregate, and nullable fields
- Sampled columns
Here some things that you can do to prepare your data for use by AI.
Create a variety of metrics
Define a variety of metrics and ground answers in these standardized metric definitions for accuracy and consistency. IBM watsonx BI uses metrics and the underlying enriched metadata, that is used to define metrics, to answer your questions and provide insights.
Add business terms
Watsonx BI comes with predefined business terms. These business terms act as metadata to enrich data assets so that AI can better understand your data and provide accurate responses to your questions.
If your organization's business terminology is different, IBM Cloud account owners and Administrators can consider adding business terms to provide additional context to watsonx BI about their organization's data. For more information, see Business terms.
Use unique names for data assets
To avoid ambiguity and confusion, use unique and descriptive names for data assets, including metrics.
Each asset and column needs a display name and description
Review the metadata enrichment results in the metric creation process to ensure each Asset and Column has a meaningful and accurate AI-generated Display name and Description.
You can also access the metadata enrichment from the Project asset tab by selecting the relevant metadata enrichment asset.
You can accept the AI-suggested name and description or click Edit to add your own.
When you edit the Display name or Description in the metadata enrichment asset, the updates might get overwritten if enrichment is run again or if a metric definition is deleted or edited and exported. To avoid this, it is recommended to make your changes in the semantic data model instead (see #2 in the section below).
When the confidence score of an AI-suggested name or description does not exceed the minimum threshold, the corresponding cell is blank. In that case, click the pencil icon next to the suggestion and provide a meaningful name or description for that asset or column.
Descriptions must be concise and reflect the purpose of the column. During query generation in a conversation, the AI uses both the identifier, name, and description (potentially with sampled data) to select the best columns to answer the question. Avoid repetitive descriptions, where every description includes the same text, especially when the repeated content is how users will frequently ask questions.
Model data in the semantic data model to help query generation
1. Add clear column identifiers
To help the LLM understand the table schema correctly and easily, you need to ensure that the column identifiers are unique and clearly worded.
Identifiers identify objects and are used to generate SQL queries in expressions.
Column identifiers should be worded in a way that aligns with how users are likely to ask questions.
For example, if an asset is about product sales or revenue, column identifier “SALES_YEAR”/”SALES_DATE” would work better than “CURRENT_YEAR”/”CURRENT_DATE”, as it describes what the temporal columns represent. Similarly, if the asset is about sales target, use column identifiers such as “SALES_TARGET_YEAR“/”SALES_TARGET_DATE”.
On the other hand, when multiple fact metrics come into play (for example, revenue + planned revenue in a single metric definition with a shared time dimension), having more general temporal column identifiers might be beneficial.
Updating column identifiers
To change the column identifier in a metric definition:
-
On the Data and Metrics tab, open the semantic data model that has the metric definition you want to edit.
-
Click Advanced mode on the Metrics overview page.
-
Under the metric definition, select the column you want to add the identifier to.
-
Open its Properties tab and scroll down to the Advanced section.
-
Enter the Identifier value.
An identifier's first character must be a letter or an underscore. Subsequent characters can be letters, numbers, or underscores, without spaces.
A Properties panel displays the Column identifier field. -
Under Actions, click Save to save the semantic data model.
-
Select the metric definition that you just made changes to and click Export metric definition. Exporting the metric definition runs metadata enrichment again and updates the existing metric, making it available for use in conversations.
2. Add column labels and descriptions
Make sure that you add a label and description for metric columns in the semantic data model. The order of priority for the label and description is the following:
-
User-defined in the semantic data model
-
User-defined in the metadata enrichment asset
-
AI-suggested or generated in the metadata enrichment asset
Writing a strong description
Ensure that the label and description are worded based on your business context and how users will likely ask questions.
- Include values in the description to specify time periods
-
For example: If months are represented as “Jan”, “Feb”, “Mar”, include that in the description. “Month values are abbreviated as Jan, Feb, Mar...”
- If values follow a known standard, mention the standard name in the description
-
For example: “Country codes follow ISO 3166 standard” or “Currency codes use ISO 4217”
- Reflect user language preferences
-
For example: If there is a column called "manager”, but your business users use the word "boss" then reflect that in the description.
-
If you want AI to use "Cost of goods sold" column to answer questions about "Cost breakdown", then add that phrase to the description of the column.
Adding column labels and descriptions in the semantic model
-
On the Data and Metrics tab, open the semantic data model that has the metric definition you want to edit.
-
Click the Advanced mode.
-
Under the metric definition, select the column you want to add a display name and description to.
-
Open its Properties tab and enter the name in the Label field and description in the Description field.
-
Under Actions, click Save to save the semantic data model.
-
Select the metric definition that you just made changes to and click Export metric definition.
3. Double check the “Usage”, “Aggregate” and “Nullable” fields for each column
In the semantic data model, check the column properties to ensure that the Usage, Aggregate, and Supports NULL values fields are correct.
-
Usage: This property applies to tables and controls how the query engine understands and processes the table, and its child objects, in a query.
-
Aggregate: This property applies to columns, and defines the type of aggregation that is applied to a summary column.
-
Supports NULL values: Specifies whether a column supports null values. By default, this property value is inherited from the source. You can change this value.
For example, for a numerical column called Revenue, the Usage field should be Measure, rather than Attribute. If the sum of the revenue is commonly used in a question, the default Aggregate should be Total, rather than Average.
Sometimes data like “unit_cost” or “unit_sale_price” is treated like a Measure. However, these are actually attributes of a product.
4. Use instructions for semi-additive measures
If you have a semi-additive measure in your data, such as headcount or inventory, which cannot always be aggregated, you can add instructions in the description directing AI to use the appropriate calculation based on the time dimension.
For example, you might have headcount data on a monthly basis but if you want the headcount for the year, you cannot add monthly headcount to get the yearly total. In this case, the value from the last month of the year would yield the correct yearly headcount.
You can include a description instructing the AI on how to calculate headcount for the year:
“Use this column for yearly headcount. The headcount for 2025 is the total headcount for December 2025.”
5. Remove unnecessary columns from the data
Remove columns from metric definitions that might not be useful in answering questions by deleting them or hiding them from the base table or metric definition.
For example, if a metric definition contains both “COMPANY_KEY” and “COMPANY_ID” columns, the AI might have difficulty deciding which one to choose when generating the query. If a business user is likely to only ask for “COMPANY_ID”, remove “COMPANY_KEY” column from the metric definition or hide it. Any columns that you hide, appear in grey and are moved to the bottom of the semantic model panel.
If you have multiple language-specific versions of the same column, such as the “PRODUCT_TYPE_” columns in the example below, you can create a new unified “PRODUCT_TYPE” column with an expression that can automatically pick the appropriate language-specific version, and hide the other redundant “PRODUCT_TYPE_” columns.
6. Data quality matters
Ensure that column values are correct and relevant.
Sampled column values are used by the LLM to construct well-formed filters. When you ask a question, watsonx BI’s NER step extracts filter values from the question, converts them into embeddings, and compares the filter embeddings with the columns’ sampled embeddings that are stored in Cloud Object Storage. The matched sampled columns feed in to the prompt to help the LLM craft appropriate filter expressions, which are used in the WHERE clause of the final query statement.
For example, if a metric definition contains a “MONTH_NAME” column, the sampled values from that column should only contain valid month names.
7. Using column expressions for abstracting complex logic
Use column expressions for calculations or transformations that encapsulate complex logic.
You can create column expressions in the semantic data model. Select a column and in the column properties, click the View or edit link in the Expression field. Add the necessary logic for this column to the Expression editor.Here is an example of an expression for the “EXECUTIVE_INDICATOR” column, which is a binary “Y” or “N” value that determined by the corresponding value in the BAND column.
Let’s say you have a "Status" column with multiple cell values like "Todo", "In progress", 'Reviewing", "Icebox" and "Done". Your organization defines an open state as a status that is not in “Done” or “Icebox”. You can create a calculation column called “Is_open” and create an expression that returns “Y/N” based on the value from “Status” column, making it easier for the LLM understand the logic.
Give your expressions column identifiers that map to how your users will reference the data. For example, the "Is_open" expression we created, is a good name if this is how users refer to open or closed items. However, if users generally refer to items as active or inactive items then "Is_active" would be a better name.
8. Using column expressions to filter data
Use column expressions to filter out unnecessary data from the metric definition before exporting it to create or update a metric.
For example, if a table includes product names in all languages and you only want the English names, you can add a filter to the table and apply it on the “product_language” column to filter out unnecessary values.
