How to Use the AI Widget

Overview

This video teaches you how to use the AI Generic Widget to create a custom visualization from a plain-language instruction instead of picking a predefined widget type. It covers attaching a model query dataset, writing an instruction for what to visualize, and using the message box to refine the result without rewriting the original prompt from scratch.

Facts to Remember

• The AI widget can generate its own dataset if none is attached, but attaching a predefined model query dataset gives you control over exactly what data it visualizes.
• The message box lets you iterate: you can type follow-up modifications there instead of editing the original prompt.
• This widget is best suited for visualization needs that fall outside Una's predefined widget types (chart, KPI, matrix, etc.).

Transcript

Let's go over how to use the AI widget.

So first you add a widget by clicking on that button and you can see at the very bottom there's an AI generic widget.

So what this allows you to do is create any widget type, not just the ones predefined in Una.

So it gives you a lot of visualization flexibility.

And if you right click at the top of the widget, you'll see a context menu.

Let's click on edit.

And you can name the widget, give it some instructions.

It can generate a data set if you don't attach one, but we're going to want to attach predefined data that we are going to set through a model query.

And then you could give a message to execute it, and then you'll see a preview once it's done loading.

And then you click the refresh button in order to actually you know load the visualization so we're going to start by adding a data set so click on the down arrow and click on model query data set and then click the edit icon so i have a data set that's predefined so i want to access the metrics table because we're going to use some finance data we're going to visualize opex by department So what this is, is so the data sits under vnew, the new measure, actual scenario, the period 2024, the department.

So here we're doing, we're using an exclude function.

So we're saying I want to see all departments, the children of them, which are all of these, except for none.

So I'm excluding none here in this function.

And then I want to pull data from operating expenses because we're visualizing OpEx by department.

So you can see a preview of that data here.

We have all the different departments and the operating expense data.

And I'm happy with that.

So I'm going to click OK.

And then now we have the data set attached.

So I can say I can I can rename the data set and say.

Topics include the Model Builder, Explorer, dimensions and hierarchies, data models, canvases and hierarchies.

Topics include the Model Builder, Explorer, dimensions and hierarchies.

Topics include the Model Builder, Explorer, dimensions and hierarchies.

And then I can use this message to just say go so that it will execute the instructions.

And the nice thing about this message box is you don't have to change to the original prompt.

If you want to modify it, you can just say the modifications here and then it will reload the widget and generate a new visualization.

So here we have a preview.

If we refresh it, you can see it and then Let's see what it generated here.

So we have all our different operating expenses.

And you can see that sales is the highest here.

And it did it in the trendline view in the color purple.

And then we could just name this operating expenses by department.

And there you have it.

You have your AI widget visualization.

Related Articles

• Widgets — the hand-built alternatives
• Dashboards — the canvas the widget lives on