Use elluminate Assist in Data Central

Permissions: All Data Central users can use elluminate Assist.

elluminate Assist is a generative AI chatbot that simplifies access to SDTM clinical trial data in Data Central by translating plain-language queries into SQL code and instantly returning relevant results.

Access elluminate Assist

Important: If this is your first time using elluminate Assist, you are unfamiliar with this feature, or you want to access the Quick Prompts, it is highly recommended that you read the General Information, Liability Disclaimer, and Tips for Effective Use. Click the Instructions button or link to access the instruction page.

Click the elluminate Assist icon in the upper-right corner of the master header. The elluminate Assist window opens, displaying links to instructions and an input field to enter a plain-language prompt.

Instructions & Information

The Instructions page has three tabs, which include the following information:

About—General Information and Liability Disclaimer: It is critical to read this before using the chatbot. A careful review of this disclaimer is highly recommended to acknowledge the functions and limitations of using AI to access SDTM data.
Tips—Best practices and practical usage guidelines categorized as follows:
- User-Specific access
- Querying and Data Specifications
- Ensuring Accuracy
- Managing Chats
- Historical Prompts and Responses
Quick Prompts—Use the pre-defined prompts to get started. Subject matter experts have compiled 27 of the most frequently asked questions for a study.

Click the 'X Close Instructions' link in the upper-left corner to close the instructions.

Ask Questions

Use pre-defined prompts or write custom ones. Examine the structure of the pre-defined prompts to better frame and optimize questions for crafting custom prompts. Once a prompt is submitted, elluminate Assist starts processing the query and generating results.

Select a Pre-defined Prompt

To access the list of Quick Prompts, click the View more prompts link from the main window. Alternatively, click the Instruction link and navigate to the Quick Prompts tab.
Select the question you wish to ask, and then click on it to send.

Write a Custom Prompt

When writing a custom prompt, it is important to note these guidelines:

Always enter the prompt as a question or direct instructions.
Keep your questions limited and focused on SDTM data in the supported domains (AE, CM, DM, DS, EX, LB, MH, VS, SV, EG, and PR). Be precise when asking your questions to avoid ambiguously phrased inquiries.
For complex questions, break them down into simpler components. elluminate Assist remembers the previous question, allowing you to ask as many follow-up questions as needed and leveraging the historical content.
For incomplete date variables, elluminate Assist assumes the first day of the month if no day is specified or January if no month is specified.
To start a new session with a new line of questions, click the plus icon in the upper right corner of the chatbot.
Questions can be up to 500 characters long.

Send a Custom Prompt

Enter the question in the input field at the bottom of the main window.
Click Send. Processing may take a few moments.

Important: elluminate Assist cannot modify or delete data. Avoid sending direct instructions to modify or delete information.

Tip: While the chatbot generates the results, hover over the user icon to display the User ID along with the date and time the question was submitted. This information remains accessible when returning to a session, confirming that the chatbot session is linked to the user.

Interpret Results

The chatbot provides a structured response consisting of:

Breakdown of SQL Steps: This describes the logical steps involved in the actual SQL query, the breakdown of the logic in plain-language, and the usefulness of the returned data.
Open result in a Listing: The results can be opened and saved as a custom listing in Data Central. Refer to the Open in Custom Listing section for details.
SQL Query: This refers to the actual SQL query being used. Click the Sources link to view the SQL Query.
Preview of Data: This displays the returned data's first few rows.

Tip: Always review the explanations and results to ensure accuracy. If needed, refine the question and request an updated response.

Open in Custom Listing

The results can be opened in a Draft Custom Listing and then saved in Data Central as a Custom Listing within the Visualizations section. Once saved, it can be docked, edited, and utilized like any other Custom Listing stored in Data Central.

Open and Save Results as a Custom Listing

To open and view a draft version of the results, click the Open result in a listing link.
A draft version of the result set opens. This is a temporary listing that can be viewed and exported to Excel. Although it cannot be docked, it can be maximized.
To save and reuse the generated listing, click the Save icon located in the upper-right corner of the listing panel.
The Save Custom Listing window opens.
Enter the Name of the Custom Listing (maximum 100 characters allowed).
Select the folder where the listing will be saved.
Select the Public or Private radio button. Only selectable with Data Central Designer privileges.
Select the Roles that will see and have access to the visualization. (This applies only to saving as public).
Update the Scope. The study you are currently in is selected by default. To make any necessary updates, click the sections Therapeutic Areas, Compounds, Programs, or Studies.
Click Save to confirm or Cancel to cancel changes.
The newly saved custom listing is located in Data Central as a Custom Listing within the Visualizations section and respond to Subject Filters and Data Cuts.

Edit Results

To edit the results, click the Edit icon next to the custom listing in the left navigation under Visualizations. The Edit icon is also located in the toolbar of a Custom Listing that is open in Data Central.
The Custom Listing Designer opens with a preview of the listing alongside the SQL query displayed in a SQL Editor.
Modify the query as needed, then click the Validate & Update Preview button. Note that if a user modifies the SQL, the Save button remains disabled until they click Validate & Update Preview.
After updating and validating the syntax and results, click Save to keep changes and exit Edit Mode.

Toolbar Actions

Custom Listings feature a toolbar in the top right corner of both draft and published listings; the designer does not have a toolbar. Below are the icons available for a Custom Listing.

ICON	ICON NAME	DESCRIPTION
	Edit	Click to open the designer window.
	Save	Click to save the draft as a Custom Listing.
	Previous Page	Navigate to the previous page in the listing. Unique to Custom Listings.
	Next Page	Navigate to the next page in the listing. Unique to Custom Listings.
	Export	Click to export the listing as an Excel file. The Excel file for the custom listing includes the data, metadata, and the export timestamp.
	Search	Easily find a column in a data or custom listing. The Search icon provides a list of all displayed column names and labels. You can type to search or scroll through the list to find the desired column.
	Dock Item	Click to dock the visualization to your sheet.
	Maximize	Click to maximize visualization.
	Restore	Click to restore the visualization to the previous size.
	Close	Click to close.

Note: When an icon is highlighted blue, it indicates that it is activated / on, whereas when it is gray, it is deactivated / off.

View Chat History

elluminate Assist stores up to 10 previous sessions in its history that the user can return to. Within each session, the last 30 days of the conversation are saved.

To access previous sessions, click the down arrow in the upper-right corner of the chatbot.

Note: The response data is not automatically updated when returning to a previous session; the question must be resubmitted to refresh the data.

Note: Each user’s chatbot is personalized and unique to their elluminate account. Data retrieval is limited to information accessible to users in compliance with their privileges and data blinding settings.

Best Practices & Tips

Review the instructions before getting started.
Begin with basic questions to familiarize yourself with elluminate Assist’s capabilities.
Experiment with phrasing to optimize query results.
Use the generated SQL to perform additional analyses and validations beyond the chatbot.
elluminate Assist is study-specific and cannot pull data from multiple studies.