Health scores in customer success will die soon because of AI!

dhiraj_patel

👉 No actual data backing while assigning weights

👉 Not updating weights with updates in product and customer behaviour

👉 It gives you a sense of health not if they are more likely to expand or churn

Here’s how this can be solved:

👉 Assign weights using AI

👉 Auto-update weights using AI to keep them updated with product changes and customer behaviour

👉 Two different pipelines for churn and expansion


Health scores will be replaced by outcome-driven signals for each specific use case, such as churn or expansion.

AI will bring long-pending disruption in customer success.

Have you adopted AI in CS already? If yes, what use cases?

Ed Powers

This assertion is way too premature, @dhiraj_patel. While it's true that many customer health scores are based on subjective criteria and "pseudo-math" and can be improved using data-driven methods and mathematical models (regression, decision trees, machine learning), stating AI is the best answer sets aside the requirements and tradeoffs in using it. The litmus test, in my experience, is improvement in predictive accuracy, given each environment's constraints. Unless AI makes a meaningful contribution, simpler and less costly methods offer better solutions.