Strategy ONE

Refine and Edit AI Datasets

Beginning in Strategy One (March 2025), bots have been enhanced to improve responses and the bot creation process. For more information on the enhancements, see Strategy One (March 2025) Enhancements.

  • You can continue to create bots, as well as edit and use bots created in previous versions. For an introduction, see Auto Bots: Customized Stand-Alone Bots (prior to March 2025).

  • To enable the new bots, contact Strategy support. For an introduction to the new bots, see Auto Bots: Customized Stand-Alone Bots. For steps to create a new bot, see Create a Bot.

Refine and edit your data when you create a new bot or AI dataset collection.

  • An AI dataset collection contains at least one AI-enabled dataset. Beginning in Strategy One (June 2025), you can also select unstructured data as datasets. You review, correct, and manipulate the datasets to refine and integrate the data. An AI dataset collection can be used in multiple auto bots, providing consistency and efficiency.

  • When you create a new bot, you can use an existing AI dataset collection or create a new one.

Refine the AI dataset collection to integrate multiple data sources; create derived metrics and, beginning in Strategy One (May 2025), derived attributes; link attributes for seamless data blending; and perform various operations on the datasets. To ensure data accuracy, create grid visualizations to validate values after data wrangling operations, ensuring the correctness of the modified datasets.

Prerequisites

  • You refine and edit AI-enabled datasets. Enabling a dataset for AI enriches it with automatically-generated descriptions of the cube itself and each column, providing the context needed for your auto bots. Use Workstation to enable datasets.

  • You can use multiple sources of data, such as two datasets, in a single AI data dataset collection. When you do, ensure a relationship between the datasets exists (for example, a shared attribute) so you can link them. The bot can then more effectively understand the relationship between the data and surface relevant information.

  • Beginning in Strategy One (June 2025), you can also select unstructured data as datasets.

    • Unstructured data includes PDF, Microsoft Word, HTML, markdown, and text files. For example, you can integrate FAQs in a Word file with troubleshooting guides in PDFs to use in a customer support bot. Another dataset collection can include PDFs of policy documents, contracts, and compliance guidelines to precisely answer legal and regulatory queries. Use Workstation to add unstructured data.

  • Beginning in Strategy One (June 2025), you can create a universal bot (multiple bots combined into a single intelligent assistant) by selecting multiple bots as the data sources.

  • To access advanced mode and perform basic data manipulations, you must have the Run AI Bots and Create and Edit AI Bots privileges. To perform advanced data wrangling, you may require additional privileges.

Create a New Bot

  1. Click Create New and select Bot.

  2. If you have access to multiple projects, from the Create Bot In drop-down list, select the project to create the bot in.

  3. You can use an existing AI dataset collection or create your own.

    • To select an existing AI dataset collection, click AI Data Collection and then select the collection(s). Do one of the following:

      • Refine and edit it by clicking Preview and then prepare and edit the datasets.

      • To use the AI dataset collection as is, click Continue and continue creating the bot beginning at this step.

    • To create a new AI dataset collection for this bot, select at least one data source. You can combine different kinds of data, such as structured and unstructured.

      • To select a structured dataset, click Structured Data, then select the dataset(s). Only AI-enabled datasets display.

      • Beginning in Strategy One (June 2025), you can select unstructured data. Click Unstructured Data, then select the file(s).

    • If you are creating a universal bot (multiple bots combined into a single intelligent assistant, available beginning in Strategy One (June 2025)), click the Bots tab and then select the bots. Click Create and create the bot beginning at this step.

  4. Click Create to create your bot with the selected data. The New AI Dataset Collection window opens.

  5. Prepare and edit the datasets.

Create a New AI Dataset Collection

  1. Click Create New and select AI Dataset Collection.

  2. If you have access to multiple projects, from the Create Bot In drop-down list, select the project to create the bot in.

  3. Select at least one data source. You can combine different kinds of data.

    • To select a structured dataset, click Structured Data, then select the dataset(s). Only AI-enabled datasets display.

    • Beginning in Strategy One (June 2025), you can select unstructured data. Click Unstructured Data, then select the file(s).

  4. Click Create. The New AI Dataset Collection window opens.

  5. Prepare and edit the datasets.

Prepare and Edit AI Datasets

  1. In the New AI Dataset Collection window, you can rename an attribute or a metric. Right-click the object in the Data panel, select Rename, and type the new name.
  2. You can change a metric's number format. In the Data panel, right-click the metric, point to Number Format, and select the type of numeric formatting (such as Fixed or Currency) to use. Define the formatting options and click OK. The formatting applies to the metric display in bots.
  3. If the AI dataset collection contains multiple datasets, link them to join attributes shared across datasets. In the Data panel, right-click an attribute to link and select Link To Other Dataset. Select the attribute to link to. Click OK. For more detailed steps and information about using attribute forms, see Link Shared Data Across Multiple Datasets.
    • Linking datasets allows them to share information based on the linked attributes and to calculate accurate metric values when data from different datasets are shown in the same visualization. It also allows the bot to more effectively understand the relationship between the data and surface relevant information.
  4. You can create derived metrics, that is, new metrics based on existing objects. Derived objects present your data in different ways. When you create a derived metric in the collection, Auto knows the formula and can describe it.
    • Create a derived metric by selecting the aggregation function (such as Sum or Average) used to calculate values in an existing metric. In the Data panel, right-click the metric to use, point to Aggregate By, and select the function.
    • Create a derived metric. In the Data panel, click Create Objects and select Metric. In the Metric Editor, create the derived metric.

  5. Beginning in Strategy One (May 2025), you can create derived attributes, that is, new attributes based on existing objects. Derived objects present your data in different ways, enhancing your data analysis. For example, you can combine the geographical region attribute with the state attribute to produce a result like Chicago, Illinois by using the Concatenation function. When you create a derived attribute in the collection, Auto knows its definition and creates its description for your bot.

    1. In the Data panel, click Create Objects and select Attribute.
    2. In the Attribute Editor, create the derived attribute.

  6. To test data accuracy, drag attributes and metrics into the Rows and Columns drop zones in the Editor panel to populate the grid visualization.

  7. To test the visualization, you can filter it. Click Filter . Create the filter by dragging attributes and metrics from the Data panel to the Filter panel. Use the filter to change what the grid visualization displays.

  8. In the top navigation pane, you can choose to Undo a previous action, Redo a previous action, Refresh the data, Pause/Resume Data Retrieval, Add Data to add more AI-enabled datasets to the collection, and add more visualizations with the Add Grid icon.

  9. When your data preparation and testing is complete, click Save in the toolbar.

    10. Changes made to the test visualization do not affect the bot’s interface. Only changes made to the dataset itself, such as editing the attribute form, formatting metrics, or creating derived attributes or metrics, are saved and reflected in the bot.

  10. Type a Name for your AI dataset collection.

  11. You can certify it by selecting the Certified AI Dataset Collection check box.

    Certifying a collection means that it has been reviewed and approved as an official source of content based on reliable data.

  12. Select the folder to save it in.

  13. Click Save.

    • If you are creating a bot, continue creating the bot beginning at this step.

    • If you are creating an AI dataset collection, close the New AI Dataset Collection window by clicking X on the toolbar.