Artificial Intelligence (AI) is a technology for tackling complex problems such as answering questions. For instance, I got the following response when asking ChatGPT, an AI powered chat bot, "Why AI":

Why AI

by ChatGPT


AI, or artificial intelligence, has the potential to revolutionize many aspects of our lives, from healthcare and transportation to entertainment and education. AI systems are capable of analyzing vast amounts of data, detecting patterns and insights that may be difficult or impossible for humans to discern. They can also automate tedious or dangerous tasks, freeing up humans to focus on more creative or fulfilling work.

AI has already made significant contributions to fields such as natural language processing, computer vision, and robotics. It has the potential to greatly improve medical diagnoses, personalized healthcare, and drug discovery. It can also help reduce energy consumption and waste in industries such as manufacturing and agriculture.

Overall, AI has the potential to create significant benefits for society, but it also raises important ethical and social concerns, such as job displacement and bias in decision-making. Therefore, it is important to approach AI development and deployment with caution and consideration for its potential impact on individuals and society as a whole.


1. Salesforce Einstein Prediction

Machine learning is a type of AI that enables computers to solve problems based on data examples. Einstein Prediction is based on an end-to-end automated machine learning engine developed by Salesforce. It is used to predict asset failure in Calem in the coming release of R2023a.

2. Calem Enterprise Lightning

Calem Enterprise Lightning is a a native application On the Salesforce platform™ for Enterprise Asset Maintenance Management (EAM/CMMS). It is available through Salesforce AppExchange since 2018. 

3. Asset Failure Prediction

Asset Failure Prediction is implemented through an integration of Calem Enterprise with Calem Enterprise Lightning on the Salesforce platform™.

  • Assets from Calem Enterprise are synchronized in the background to the Calem Enterprise Lightning in the Salesforce.
  • Asset failure prediction is calculated in the Calem Enterprise Lightning.

4. Asset Data Segmentation

Asset data is segmented based on Asset Failure status which is available from R2023a. Example data is prepared by data analysts of Calem customers so the prediction is trained for the customer assets.

  • At least 400 assets are required for failure prediction
  • At least 100 assets that failed within lifespan. 
  • At least 100 assets that have not failed within lifespan.
Asset Fail Status Data Segmentation
Failed These assets failed within lifespan. They are used as examples for failed assets.
Not-FailedThese assets do not fail within lifespan. They are used as examples for assets not failed.
PendingThese assets will be predicted for failure.
Not-ApplicableIgnore asset failure prediction for assets of this fail status.

 5. Einstein Prediction Builder

Go to Admin module of Calem Enterprise Lightning and search for "prediction". Einstein Prediction Builder will show. Launch the builder to create a prediction for asset failure. 

  • The prediction type is "Yes/No" for predicting the probability of an asset failure.
  •  Select Calem asset object for prediction: Asset (EAM).
  •  Do not include assets of fail status "Not-Applicable" for prediction.
  •  Define the Yes examples that asset failed.
  •  Define No examples that asset do not fail.
  • Select or exclude fields to be included in asset fail prediction. All fields in the Calem Enterprise Lightning are selected for prediction. 
  • Predict all asset for failure excepting example assets.
  •  Add a new custom field to store prediction scores.

That is all to create the asset fail prediction. It takes a while for Einstein to provision the prediction. You may come back to the prediction builder later to continue the prediction setup.

  • Once the prediction is processed its scorecard is available.
  • See this article to understand the scorecard.
  • Next, enable the prediction.
  • It may take a while for Einstein to generate asset failure scores.

6. Evaluate Asset Fail Prediction

The evaluation of asset fail prediction help understand the prediction including answers for the questions below. 

  • How effective is the asset fail prediction?
  • What is the prediction score range that indicates an asset is likely to fail?
    • For instance, an asset of prediction score of 54 or above is likely to fail soon. 
    • Business processes and alerts can then be configured based on the evaluation.
  • See this article for steps to perform the evaluation.

The evaluation may be started after the prediction is enabled, and there are assets with prediction scores failed or not failed within lifespan. These assets are necessary for comparing prediction scores against realities.