Whitepaper · Balboa × Pendo
Predictive Customer Success
Building the data foundation for an agentic CS motion.
Nearly half of CSMs miss their annual target. When 44% of the role falls short, the problem is the architecture, not the people. The signals that predict renewal, expansion, and churn already exist across product analytics, the CRM, support, and billing — today, humans are asked to synthesize them by hand.
This joint field guide from Balboa and Pendo walks through the shift from reactive, dashboard-driven customer success to predictive, agentic customer success: which data sources to use, how to make them model-ready, how the model works, how predictions become action, and who you need to run it.
No email address required
Published July 2026
The core argument
The CRM was never built for analytics. Customer success platforms tried to fill that gap and created a shadow CRM. Meanwhile the clearest signal — product usage — sat in the product org, siloed away from the teams responsible for retention. So the wrong kind of work fell on humans: CSMs synthesizing a health score, a last-login date, and a renewal date by intuition.
A better system is now within reach for any company with a product and a customer base. A predictive model ingests every signal, learns from your own history of churn and renewal, and produces one continuously updated answer for each account. That answer lands in the workflow your team already uses, explained in plain language, paired with a recommended next step, and acted on for you.
You do not need a data science team, a platform migration, or a year of build time. You need the data you already have, the right architecture, and a few strong people who know what they are doing.
The shift, in one table
No email address required
Ready to build the data foundation?
We help enterprise CS, product, and RevOps teams design and stand up this architecture on Pendo and the systems they already run.