Design an ML App that predicts, classifies, recommends, or decides options based on your use case.

Step 1: Create an ML App

Get started by clicking on the Create ML App Icon in the top left corner, and then give your ML App a name and description. Let's call this ML App "Suggest Product Category" and give it the description "Suggest recommended product categories for purchasing on the Order Review page."

When you're ready to move to the next step, click on Create ML App.

Step 2: Decision Options

In the example, the ML App is deciding the order to display the product categories on an order review page of a checkout flow. The product categories are the decision options: Produce, Deli, Canned Goods, or Alcohol. The Optional ID can be used for an internal code, if left blank it will auto-generate.

When ready, select the "Next" button in the top right corner of your screen.

Step 3: Goals & Data

Next, give the ML App its Goals to work towards. Goals help an ML App learn what it should do. In the example, the ML App's Goals are to:

Decrease OrderReview_Abandonment
Increase CartValue.

Tip: Goals are either a number or currency.

Add Data that you think might influence the decision made by the user. You know what makes your users tick, this is where that expertise will be very helpful.

Some data that might influence the ML App is:

LastPurchaseItem
LastPurchaseCategory
AgeInYears
AvgOrderValue
TimeofDay

(Optional) Add Guardrails

Do you want to add some Guardrails to the ML App? Hard and fast rules are easy to include.

Step 4: Save the ML App

When you're done designing your ML App you can save it. When you save your ML App, SAVVI will run Health Checks to ensure the AI framework you've designed is set up correctly and suggest improvement opportunities wherever necessary.

Step 5: Publish the ML App

Select the "Save & Publish" button when designing the ML App; or
Submit the ML App for publishing from the list of drafted ML Apps

Your ML App is ready to be deployed in your software! 🎉

Next, you may: