Flows are self-contained workflows within the Virtualitics AI Platform. Flows are modular, highly customizable, support a wide array of analytics routines, and can be shared across users with access to a deployed instance of the Virtualitics AI Platform. A single instance of the Virtualitics AI Platform may contain one or more Flows.

Three Flows displayed on the Home page of an instance of the Virtualitics AI Platform.

Why are Flows useful? When would I use Flows?

Flows are at the heart of the Virtualitics AI Platform - anything you do within the Virtualitics AI Platform outside of Explore will require a Flow.

Flows would be useful in the following scenarios:

• You need to programmatically access data for an analytic workflow.
• You have a repeatable analytic workflow that you wish to automate.
• You have a machine learning pipeline that you would like to run without having to expose code.
• You have a machine learning model that you would like to interrogate.
  • Refer to the knowledge-base articles How to use the XAI module and How to use the SPT (Scenario Planning Tool) to learn more about Explainable AI (XAI) tools available in the platform.

How do I use Flows in the Virtualitics AI Platform?

Flows may be made available in the Virtualitics AI Platform in one of two ways:

1. Developers may create new Flows and add them to the Virtualitics AI Platform using the “Create Flow” button on the Home page.
   • To learn how to create a new Flow, please refer to this set of knowledge base articles.
2. If Virtualitics (or your company’s) data scientists developed Flows that were available when the Virtualitics AI Platform was deployed (or were subsequently integrated into your instance of Predict), those Flows will be accessible from the Home page.

From the Home page, a Flow may be run by:

1. Clicking on the tile corresponding to the Flow, 
2. Clicking on the prompt to start the Flow (“Start Flow”), and 
3. Proceeding through the Flow by clicking the “Next” button at the bottom of each page.

Each time you run a Flow, a new row will appear in the table on the “Flows” page. From the “Flows” page, you can (1) set up a time for a Flow to run on a schedule, (2) return to a specific point within a Flow if interrupted, (3) favorite/share Flows, and (4) delete Flows. In the example shown above, a user has selected the “Predicting Gearbox Failure” Flow and run it through the “Getting Predictions” Step.