Skip to main content

Documentation Index

Fetch the complete documentation index at: https://help.maestra.io/llms.txt

Use this file to discover all available pages before exploring further.

​
Why is the algorithm useful?

— It identifies customers who are likely to make a purchase following a personalized offer (personalized product recommendations) and automatically sends them relevant emails. — It maximizes the open and conversion rates to make up to 10x revenue compared to non-ML algorithms, which periodically send emails after an order.

​
How to set up a flow

​
1. Launch the algorithm.

Go to Campaigns → Next Best Action: 00_nba After launching the “Next Best Action” algorithm module will be in a demo mode so it can only be activated up to 10 times per day. To access all features without restrictions, contact your manager to enable the module.

​
2. Create the flow.

2.1. Select the Time for Next Best Action event that would trigger this flow: 01_nba
To make the event accessible, launch the algorithm at least in the demo mode.
2.2. Add the Delay block to set the delay time before the next step. You may also configure node schedule and limit sending hours: 02_nba 2.3. Then add your specific conditions. Don’t forget to check if customers are subscribed with valid contact details: 03_nba Make sure the contact details and the subscription you are checking in the condition block match your communication channel. 2.4. Send an email: 04_nba 2.5. You can limit how often the flow should run for each customer: 05_nba 2.6. Your flow is ready, now you can launch it: 06_nba

​
How the algorithm works

The algorithm is powered by a neural network. It analyzes the following customer actions:
  1. Purchasing a cheap product;
  2. Purchasing a medium-price product;
  3. Purchasing an expensive product;
  4. Viewing a cheap product;
  5. Viewing a medium-price product;
  6. Viewing an expensive product;
  7. Receiving an email;
  8. Opening an email;
  9. Clicking on an email links.
The neural network also transmits intervals between such actions. The target variable is the last 5 customer actions. They are what the neural network learns to predict. The output AI model shall predict whether it is reasonable to send an email this day. If the model decides to send it, the Next Best Action flow is triggered.