The only academically grounded framework for CS AI decisions. Developed from 15+ years of enterprise CS delivery across EMEA and APAC, and doctoral research at the University of Canberra on predictive churn modelling, CS ambidexterity, and service-dominant logic.
The state in which a customer organisation has internally disengaged from a vendor's platform — reduced internal advocacy, stopped integrating workflows, begun the psychological process of churning — while maintaining the surface appearance of an active account. It is called "institutional" because it is a collective organisational shift, not a single person going quiet. By the time you notice it, you are typically 60–90 days behind the decision.
The systemic tendency of CS teams to rely on indicators that surface churn risk only after the customer's internal decision to leave has effectively already been made. The metric moves. The CSM reacts. The account is already lost. The Latency Trap is not a technology failure — it is an architectural one. It is what happens when a CS function is designed around response rather than prediction.
A metric measuring the degree to which a customer's core business workflows are embedded within your platform — not just whether they are logging in, but whether they are dependent on your platform for operations that matter. High Integration Density accounts have genuine workflow embedding and high switching costs. Low Integration Density accounts are at risk regardless of what their health score shows.
The term for the financial exposure created by Institutional Silence in any given account or across a portfolio. It is the most actionable number you can generate from your CS data — more useful than NPS, more honest than a health score, and more specific than churn rate. For a portfolio of 80 accounts, the SuccessEngine™ calculates the full Silence Tax exposure in approximately 90 seconds.
A systematic audit of a CS function's current operating motion, designed to answer one question: where is this team spending human attention on tasks that could be automated, and where are they using automation for interactions that require a human hand?
The Review examines six process domains: onboarding and activation, health monitoring and churn prediction, proactive engagement, escalation and save management, expansion and upsell identification, and QBR and executive relationship management.
Maps each identified process against the Automate-Augment-Anchor decision standard — the core intellectual contribution of the framework.
Automate: Repetitive, data-driven, low-relationship-risk processes. AI runs end-to-end. Humans review exceptions only.
Augment: Relationship-present but scale-limited processes. AI drafts, scores, and surfaces — the human decides and delivers.
Anchor: High-stakes, trust-dependent, irreplaceable human moments. AI supports context only — never acts.
Deploys targeted automation only to processes confirmed appropriate in the Refine stage. This is an argument for precision automation — not maximum automation — deploying machine intelligence exactly where it creates leverage without destroying relationship value.
In practice, the Automate stage involves four categories: monitoring automation, communication automation, intelligence automation, and administrative automation.
A practitioner's guide to knowing exactly which Customer Success processes to automate, which to protect, and which should never run without a human hand. 7,000 words. Grounded in doctoral research. Free to download.
Download White Paper →The free 48-hour audit applies the Review stage of the framework to your account portfolio. Your Silence Tax number is the first output. The debrief is the first conversation about what it means.
Book Free Audit →