I want to say something that will be uncomfortable for a lot of people in customer success.
The traditional high-touch model is not sustainable. And AI is making the unsustainability visible in ways it wasn't before.
What the Traditional Model Was Built For
The high-touch customer success model made sense when:
- The ratio of accounts per CSM was manageable enough that genuine relationship investment was possible
- Customer data was hard to aggregate and synthesize, so the human relationship was the primary signal of account health
- The value of each interaction came primarily from the relationship, not from the insight the CS professional brought to it
- The main risk customers needed help navigating was technical — getting the product to work
All of these conditions have changed.
What Has Changed
Account ratios have increased. The unit economics of a dedicated CSM for every account at every tier don't work as the market matures.
Customer data is now aggregatable and synthesizable in ways that produce better account health signals than the human relationship ever could. AI can tell you which customers are at risk based on behavioral signals before the customer says anything.
The value of a customer interaction now comes almost entirely from the insight the CS professional brings — not from the relationship maintenance itself. Customers who want relationship maintenance can get that from an AI-powered chat interface. What they cannot get from AI is genuine expert judgment about their specific situation.
The main risk customers face is not technical anymore. It's organizational — will they actually adopt the platform, change the workflows, develop the internal competency? That's an organizational change challenge, not a technical support challenge.
What Has to Replace It
The model that works in this environment is not "more CSMs" or "better QBR templates." It's a fundamentally different approach.
Behavioral signals, not relationship signals. Know which customers are at risk and which are ready to expand based on what they're doing, not what they're saying.
Insight-led interactions, not calendar-led interactions. Show up when you have something specific and valuable to say — not because a QBR is scheduled.
Scalable digital touch for routine moments. The renewal reminder, the feature announcement, the onboarding check-in — these do not require human interaction. Automate them well, and free the human time for interactions that actually require human judgment.
Human expertise for the moments that matter. When a customer is at an adoption inflection point, when a strategic decision needs to be made, when the organizational change is the real challenge — that's where human expertise earns its cost.
The future of customer success is not more of the same at higher efficiency. It's a different model, built for a different set of conditions.
The companies that figure this out first will have a structural cost and quality advantage. The companies that don't will spend the next five years defending a model that is slowly becoming uneconomic.
Views are personal.