AI Readiness Score: Enhancing Responses in AI Modules

Lumenore now introduces the AI Readiness Score within the Data Dictionary, empowering users to assess and enhance the quality of their data for AI-driven processes. Upon accessing the Data Dictionary, users will see the AI Readiness Score tab, offering a clear view of how their data is prepared for use in AI dashboards, AI insights boards, and Lumenore Ask Me. The score is color-coded on a scale of 100, providing an intuitive gauge of the data’s AI readiness. Users can also access below mentioned sections of the score  where they can make the necessary improvements.

AI Readiness Score Composition

The overall score is derived from weighted parameters that define AI usability:

• Business Context – 40%
• Description – 30%
• Schema Prompt – 20%
• Synonyms – 5%
• Unit of Measure – 5%

This weighting ensures that both technical structure and business meaning are captured for effective AI interpretation.

AI readiness focuses on enhancing the quality of metadata and context for AI, rather than improving data quality through tasks like handling missing values or removing duplicates.

A universal button labeled “Improve AI score” is available, allowing users to trigger automatic enhancements that optimize their data for AI use. This feature enables effortless refinement of data to ensure it meets the required standards for AI-driven insights.

Previously, users lacked an efficient way to evaluate their data’s readiness for AI-driven insights, which led to inefficiencies in data preparation and less contextual AI responses. Now, with the AI Readiness Score, users can immediately assess how ready their data is, receive automatic improvement suggestions, and optimize the data for better AI performance. This ensures more accurate insights and enhances overall platform performance.

To Access the AI Readiness Score:

The AI Readiness Score will be prominently displayed in the following places:

• Ask Me:
  • Click on “Ask Me” and then select the schema name. This will show the “Manage Schema” option, which, when clicked, will take the user to the AI Readiness page.

• • Alternatively, you can click on “Switch Schema” and, next to the schema name, you’ll find the “Schema Manager” option.
• AI Dashboard and AI Insights Board
  • Click on “Dashboard,” then select “Create AI Dashboard.”
  • Next, choose “Use Saved Schema” and click on “Manage Schema” next to the desired schema name.
  • The AI score will be displayed next to the schema name.

1. Accessing the AI Readiness Score

When a user navigates to the Schema Manager, they will find a dedicated tab for the AI Readiness Score. This score reflects the quality of metadata and the contextual information associated with data columns, which improves AI’s ability to understand and generate accurate responses.

2. Improving Your AI Score

The user has two options: they can either improve the AI score automatically with the help of AI or manually.:

• Enhance the readiness score using AI assistance:
  To improve the AI Readiness Score, users will have access to a tab labeled “Improve AI Score”. These automated enhancements will improve the score.

The following changes will occur automatically:

• • Add Descriptions: Adding detailed descriptions to data columns.
  • Add Synonyms: Enhancing data attributes with synonyms to improve contextual understanding. The user can also “Generate AI synonyms” by clicking.
  • Add Units of Measurement: Assigning appropriate units to data measures for accurate representation.

Instead of improving all components at once, users can update the AI score for an individual component by selecting the “Improve AI Score” icon positioned at the top right of each section.

The Review option at the top right corner of all components allows users to quickly navigate to the exact place where they can complete or update the pending items related to AI readiness.

Users can also enhance their metadata using built-in AI tools. In the Data Dictionary tab, Generate AI Descriptions automatically creates clear, human-friendly column descriptions. In the Synonyms tab, Generate AI Synonyms produces relevant alternative terms based on column names and context.

• Manually enhance the AI readiness score:
  • Add Descriptions Manually: The user can manually edit or add descriptions to give more context to your columns.
  • Follow the guidelines in the ‘Best Practices for Writing Column Descriptions’ section of the Data Dictionary. Click here for more details.
    • Review and edit the Descriptions.
      Review and edit the AI-generated descriptions to ensure they are accurate, clear, and relevant to your data.

Follow:

• • • • Review and refine if required (e.g., for Customer ID: “Unique identifier for each customer, generated at registration”).
      • Use simple language—imagine explaining it to a new team member.

Example:

• • • • Column:Discount Rate (%)
      • Description:“Percentage discount applied to orders for loyalty program members. Ranges from 5% to 20%.”
    • Using Clear Column Names
      Confusing names makes data hard to understand. Clearing names saves time and reduces errors.
      Follow:
      • Avoid abbreviations or codes(e.g., cust_id → Customer ID).
      • Use full wordsthat describe the data (e.g., prod_nm → Product Name).
      • Be specific(e.g., Sales→ Total Sales).

      Example:

      • Don’t use: amt, dt, loc
      • Do use: Order Amount ($), Order Date, Delivery Location
  • Add Synonyms Manually: Enhance AI’s contextual understanding by adding synonyms that reflect your business context.
    • Mention Synonyms and Alternative Terms: In the column description, you can include synonyms or alternative terms that users may use when asking questions. This ensures that the AI can recognize different phrasings and still retrieve the correct column.Example:
      • Column Name: PROFIT
      • Description: Use this PROFIT column when the user asks about ‘profits’, ‘earnings’, revenue gains’, or ‘margins’.
      • Synonyms: earnings’, ‘revenue gains’, ‘margins’.

      Use Case:

      • User asks: “What are the total earnings for this quarter?”
      • Response: The AI should use the PROFIT column and show a chart for Profit by quarter.
  • Add Units of Measurement: Assign appropriate units to data measures to ensure accurate representation.
    • Set Unit of Measurement and Placement
      Units help make chart labels and summaries more meaningful and consistent.
      Follow:
      • Add $ for Sales, % for Discount
      • Choose Prefix or Suffix placement for units (e.g., $500 vs. 500 USD)

      Example:

      • Unit of Measure→ set symbol like $, %, Kg
      • Unit Placement→ select Prefix or Suffix
  • Schema Prompt: It boosts the AI readiness score by helping users define and structure their data schema, ensuring it meets AI requirements for enhanced understanding and more precise insights. The schema prompt contributes 20% to the total score.
  • Business Context: It refers to manually filing certain aspects of the data in a way that aligns with your business requirements, which significantly impacts the AI readiness score. It contributes 40% to the total score.
  • Here’s a breakdown of the checklist items:

a. Column availability managed as per requirement: Ensure all necessary columns are present and complete for analysis.
Too many visible columns overwhelm users. Hide irrelevant or unused fields.

Follow:

• • • Turn “on” for columns users need daily (e.g., Order Total, Customer Email).
    • Turn “off”for outdated, duplicate, or internal-use columns (e.g., Legacy ID, Raw Data JSON).

b. Column category marked correctly as dimension, measure.
Proper classification helps users analyze data correctly (e.g., defining column category as measure, dimension, or date).

Follow:

• • • Measure: Numbers you can calculate (e.g., Revenue, Quantity Sold).
    • Dimensions: Categories for grouping (e.g., Region, Product Category).
    • Date: Time-based fields (e.g., Order Date, Shipment Date).

c. Column aggregation is selected as per requirement: Apply appropriate aggregation (sum, average, etc.) based on business rules.

Follow:

• • • Use Sum for Sales, Quantity, or Profit

Example:
In the Aggregation column, choose:

• • • Sum, Avg, Count, Max, Min, or NA (if not applicable)

d. Date column marked as ‘is date’
Any column with a datetime stamp should have the ‘is date’ option set to “On.”

Example:

• • • Column Name: ORDER_DATE
    • Description: This column refers to the date when the order was placed. When the user refers to date, always use this ORDER_DATE column.
    • Synonyms: “Order Timestamp”, “Transaction Date”

e. Correct date preference is selected.
This ensures Lumenore Ask Me and Dashboards know which date field to use by default when processing time-based queries (e.g., “Show last year’s sales”).

Follow:

• • • In the Preferencestab, select the most relevant field (e.g., Order_Date)
    • This becomes the anchor for filters, trending metrics, and comparisons across modules

Note: AI Dashboards and Ask Me to auto-apply this date for time slicing, saving effort, and avoiding confusion.

Publishing and AI Score Suggestions

After making improvements using the “Improve AI Score” button, users can press the Publish button. Upon publishing, a pop-up message will appear to inform the user that their AI score will now reflect the enhancements that have been applied. The pop-up will also include suggestions on how to improve the AI score further.

Limitation

Even though Lumenore makes data prep easy, there are a few things to be careful about:

• AI Score for Shared Schemas
  For shared schemas, users will not have the ability to edit or improve the AI Data Readiness Score. This restriction ensures that shared schemas remain consistent across all users.

• AI-Generated Descriptions Might Be Wrong
  AI can auto-create descriptions for your data, but it’s not perfect. It may misunderstand the context or use generic terms that don’t align with your business.

Example:

• • • AI might label the “Customer Age” column even if the data shows account numbers.
    • Fix:Always review and edit AI suggestions to add specific details.
• Hiding Columns Doesn’t Delete Them
  The “toggle off” feature only hides columns from the user interface—data is still stored in the backend. If too many unused columns exist, this can cause confusion or slow performance.

Example:

• • • Toggling off the “Old Promotion Code” hides it from view, but it’s still taking up storage space.
    • Fix: Archive or delete truly unnecessary columns (if allowed).

• Misclassified Data Types Break Reports
  If you label a column as the wrong type (e.g., a number column as text), filters, calculations, or charts might fail.

Example:

• • • A “Revenue” column classified as text won’t let you calculate totals.
    • Fix:Double-check data types during setup (e.g., numbers = “Measure,” dates = “Date”).

• Too Many Active Columns Clutter the Screen
  Enabling too many columns makes the interface messy and hard to navigate. Users might miss essential fields.

Example:

• • • 50+ columns labeled “on” can overwhelm a new user.
    • Fix:Toggle off less-used columns and group related fields.

Note:

• LQL variables cannot be executed in Ask Me analysis. Please create SQL variables for this purpose instead.
• Variables created in the dashboard have a local scope and cannot be used in other modules like Ask Me, DYK, AI Dashboard, or the Insights Board.