Start a conversation

How to use Sheldon's AI Powered Smart Recommendations (Early Access, Beta-Trial)


Sheldon is an Artificial Intelligent (AI) system that learns from the behavioural patterns of previously churned and retained customers and helps you to understand their behaviour by providing meaningful insights. Sheldon provides this information in the form of recommendations and anomalies. Sheldon also uses advanced techniques to understand the correlation between product features, CSM activity and retention. This is what enables Sheldon to generate recommendations on the most relevant product features and CSM activities.

Who is this for?

If you are a Customer Success Manager (CSM) handling a portfolio of accounts, this article will help you use understand how you can leverage Sheldon's AI Powered Smart Recommendations to drive retention.

How can I view my recommendations?

There are two places you can view the recommendations made by Sheldon:

  1. My Tasks: My Tasks->My AI recommendations

  2. Account 360 Page: My Customers->Accounts

How to use recommendations made by Sheldon?

On individual account page, click on Sheldon's Icon to see a list of all your active recommendations. Click on "See Details" to see additional information from Sheldon.

In this view, you will see the exact details about the recommendations Sheldon makes. You will be able to see the exact correlation of the Feature with the probability of Churn.

Can I trust Sheldon recommendations?

Yes, you can trust Sheldon's recommendations as they are generated by information gathered from the behavioural pattern of the churned and retained customer from your historical data. Also, Sheldon is smart enough to understand which features are important to improve LifeTime Value(LTV) or  to improve retention.

What happens if my customer behaviour changes?

The thing that makes Sheldon from good to great is the capability to learn and unlearn.  Sheldon constantly trains himself to learn the new behavioural patterns from the customers and unlearn the old behavioural pattern(if required). This helps Sheldon to understand and capture the change in customer behavioural and act on it respectively.

What happens when we release new features in the product?

Sheldon automatically identifies new features if available and waits till enough data gets simulated after that it tries to understand the influence of that feature against other features and retention. If Sheldon finds pieces of evidence that this feature can improve LTV(LifeTime Value) or retention it will recommend that feature with a certain number that can improve  LTV or retention.

How is potential LTV Calculated?

In order to calculate LTV, Sheldon first calculates churn probability with the help of historical data, and then it uses this formula to calculate LTV  

LTV =  MRR / Churn Probability

How frequently will these recommendations be generated? 

These documents/recommendations usually get generated once a week. Typically it will be available on Monday. Also, The frequency for the recommendations can be changed based upon the request. 

How can I provide feedback and improve my AI model to perform even better?

  1. There is a dedicated section below every feature recommendation for feedback. When you click on see details for any recommendation, you'll be able to see the details for that recommendation along with you'll see a thumbs up and thumbs down for feedback. In case if you don't like the recommendation or it seems irrelevant to you, Sheldon will ask you to select a reason for that. By providing feedback you'll make the AI model even more powerful and better.

    Note: In case you find a recommendation irrelevant, In the feedback section, you can always report it. This will help Sheldon to learn from it and give you more accurate advice based upon the features. Also from next time, Sheldon will avoid that feature toward recommendations and try to pick other features.

  2. Sheldon is an AI system that learns from data and as of now he is young and the best way to help him is to provide feedback. Sheldon will take action on your feedback and try to overcome its flows. In the end, you have a refined AI System that can help you to maximize retention.

Can I hide or stop all or certain recommendations?

Sheldon gets smarter with you feedback. So we encourage you to pass the feedback. For now there is no other way to control the recommendations. But we will certainly appreciate any other suggestions around this. Please pass those suggestions over to your Customer Success Manager from CustomerSuccessBox.

Will recommendations be generated for all accounts or specific accounts?

Sheldon generates recommendations only for those accounts that have either high churn probability or low product stickiness and Sheldon's try to improve their product adoption which results in high LTV.

What happens if I take action on it?

If you take action on Sheldon's recommendations, it will reduce customer churn probability and will increase product stickiness as Sheldon provides recommendations on the features that have a good relation with retention as well as have a high correlation with other features.

What happens if I ignore recommendations?

Sheldon is very smart when it comes to identifying poor accounts. Sheldon performs several statistical calculations and uses advanced machine learning techniques to pick those accounts that are more likely to get churned in the future. Hence, If you tries to ignore Sheldon's recommendation it may result in the form of churned accounts or lower Life Time Value (LTV).

Can I enable a notification for a new recommendation?

Recommendations are generated weekly and you should visit My Tasks > My AI recommendations for the new recommendation for that week. There is not additional notification that is sent as of now, but we do plan to add it in the future.

Where can I see all AI recommendations?

Recommendations are generated weekly and you should visit My Tasks > My AI recommendations for the new recommendation for that week. There is not additional notification that is sent as of now, but we do plan to add it in the future.

Can I filter recommendations?

Yes! You can filter your recommendations based on the following attributes:

  1. Week of Generation of Recommendation
  2. Account Segments
  3. Account Name
  4. Type of recommendations (Product Adoption, Engagement, Anomaly) 
  5. Tags (Upsell or Risk)
  6. Success Roles
  7. CSM(s)

Related Articles

Contact Us

In most cases, you would have a dedicated Customer Success Manager who will be your point-of-contact for any queries or assistance. You may also,

  1. Initiate a chat directly from within CustomerSuccessBox. Use chat option from the Chat icon in the left side navigation bar
  2. Create a conversation from here
  3. Send an email to
Choose files or drag and drop files
Was this article helpful?
  1. Deepshikha Chakraborty

  2. Posted
  3. Updated