Generative AI Integration

You can use Pyramid's generative AI features to generate scripts and images, dynamic text and infographics, calculations and lists, schedules, colors, scripts, and more. This functionality makes use of an underlying Large Language Model (LLM) and allows you to generate a range of content, including complex code, simply by providing a text prompt that describes what you want in natural language.

Note: You can only make use of this functionality if your administrator has selected the Enable Generative AI checkbox in the AI Settings page of the Admin Console. If your administrator does not select this option, the function is not available for use.

A core tenet of Decision Intelligence is AI and augmented analytics, which enables users to implement predictive and even prescriptive analytics. A variety of Pyramid's core functionalities are driven by AI, optimizing the user's experience from end-to-end. This includes tools for heuristics, automated data mining, predictive analysis, deep learning, and ultimately AI in the context of data and its analysis.

In addition, AI is integrated throughout the Pyramid platform to enhance data modeling, data science, storyboard and publication design, and decision modeling. Pyramid facilitates the use of a number of LLM engines (such as OpenAI's ChatGPT engine) for a range of functions, including auto generating scripts, dynamic text, colors, and images.

Warning: When using LLMs, your assets are generated using public domain algorithms. This can produce erroneous and inconsistent or random results. Use at your own risk.

Where is AI Integrated?

Pyramid Apps

Model

You can use the Generate your script dialog to generate SQL, DAX, and MDX code for complex data extraction queries in data modeling, and to generate R and Python ML scripts that can be used to produce ML models. You can also use the Explain Code feature to generate an explanation of your scripts (whether or not the script is generated using AI).

You can use AI to generate categories (such as date-time, geospatial, or machine learning) for columns in your model that have not been explicitly assigned a category.

You can also add Gen-AI nodes to your data flow that automatically translate data, asses its sentiment, and transform it in response to your general text prompts.

Discover

You can use the LLM PQL Function to produce dynamic actions and alert schedules.

You can also use the Chatbot to build discoveries at design time using a natural language prompt, and to interrogate discoveries when opened in the Viewer. For more information, see Chatbot.

Formulate

You can use the LLM PQL Function to build custom Members (formulas), Lists, KPIs, and Scripts. The LLM function lets you generate scripts based on a given prompt when working in Script Mode.

You can also use the Chatbot to create calculations (formulas, lists, or columns) using a natural language prompt. For more information, see Chatbot.

Tabulate

You can generate scripts to build spreadsheet formulas directly into your business models in Tabulate.

Present and Publish

You can use the AI Image Generator to create images based on text prompts. This can help you to quickly and easily design your presentations and publications and add custom graphics.

You can use Auto Translate to automatically translate dynamic text into the user's application language when they open the presentation or publication.

You can also use the LLM PQL Function to produce dynamic text, placeholders, and tooltips to dynamically display AI-generated text. In Presentations, you can incorporate this function into alert schedules, tooltips , actions, and jump actions.

In Publications, you can also use the LLM PQL Function to define triggered scheduling, as well as report and page triggers.

You can also use the Chatbot to build presentations and publications at design time using a natural language prompt, and to interrogate presentations at runtime. For more information, see Chatbot.

Illustrate

  • You can use the LLM PQL Function to produce dynamic text, placeholders, tooltips, and infographic illustrations, which can then be saved, shared, and used in presentations and publications.

    • Common Interfaces

    Color Picker

    Use the Find a Color feature to generate a color using Pyramid's generative AI integration. The Color Picker is used in several locations across Discover, Present, Publish, Illustrate, and Tabulate. The Color Picker is used to change the background color of reports and charts, and to change font colors for titles, labels, and more.

    Scheduling

    Pyramid facilitates the scheduling and distribution of a number of tasks; these schedules can be configured based on PQL statements. The LLM PQL Function can be used to build scripts throughout the Pyramid platform for triggered scheduling, for alerts, publications and on-demand schedules

    Generate Your Script dialog

    The Generate your script dialog is used to generate SQL, DAX, and MDX code for complex data extraction queries in data modeling, and to generate R and Python ML scripts that can be used to produce ML models. For more information, see Generate Your Script.

    Explain Script

    Use the Explain Script function to produce an AI-generated explanation of what the script does (note this is available even if the script was not itself AI-generated). Note: Each time you click the Explain Script button, a new explanation is generated. For more information, see Explain Script.

    Chatbot

    Use the Chatbot in Discover, Formulate, Present, and Publish to build discoveries, calculations, presentations, and publications at design time or to interact with presentations at run time or discoveries in the Viewer. For more information, see Chatbot Functions.

    Using the LLM through PQL

    Important: The LLM function was formerly called the ChatGPT PQL Function, and ChatGpt remains as an alias.

    You can use the LLM PQL Function from the Common PQL functions when writing PQL expressions to use the LLM when generating various custom and dynamic assets. This functionality is supported in a number of locations, including:

    Note: You can use the LLM function to add additional information and background; the LLM cannot tell you about your data set.

    Note: The Pyramid Query Language (PQL), pronounced "Prequel," is a language built into Pyramid's PYRANA engine to allow users to construct queries that can run against the many SQL data stacks that Pyramid can query natively. PQL includes a vast array of operations and functions that allow users to query data and build analytical logic.