When a team posts a strong customer satisfaction score, the instinct is to credit the people — "we have great reps," "she has a real gift with customers." Sometimes that's true. But if a single team sustains 90%+ CSAT across a dozen people, month after month, across a rotating mix of issues and personalities, that's not a talent story anymore. Talent doesn't scale that consistently. Systems do.
I lead a team of twelve at IKEA, and we've held a 90%+ satisfaction rate through a lot of turnover, training cycles, and difficult-customer days that should, by the talent theory, have dragged the number down regularly. They haven't, because the number was never actually riding on any individual rep's personality. It's riding on a set of specific, repeatable behaviors that almost anyone can learn to execute well.
What "systems" actually means here
It means breaking a vague goal — "make customers happy" — into specific, teachable moments that consistently produce that outcome. A few examples from how my team actually operates:
Escalation handling has a structure, not a vibe. Every rep is trained on the same sequence: confirm understanding before offering a solution, set a clear expectation for next steps, follow up at a defined time. That sequence doesn't depend on the rep being naturally calm under pressure. It gives someone who isn't naturally calm under pressure a script that produces a calm-sounding outcome anyway.
Role-play happens before the real conversation, not instead of feedback after it. The team practices the actual difficult conversation patterns — the angry customer, the unclear request, the repeat complaint — in a low-stakes setting first. By the time a rep faces the live version, it's not a novel situation. It's a rehearsed one.
Coaching is specific enough to practice, not just specific enough to log. "Improve your tone" isn't a coachable instruction. "In your last call, you used the word 'unfortunately' three times in the first minute — try leading with what you can do instead of what you can't" is. The second version is the difference between a number going up and a number staying flat.
If your customer satisfaction score depends on which reps are working that day, you don't have a customer experience strategy. You have a staffing schedule with extra steps.
Where this connects to everything else
This is the same instinct that shaped The JetLyfe's customer journey — designing specific touchpoints (a follow-up message at a particular moment, a return policy explained a particular way) rather than hoping the brand's "vibe" would carry the experience. And it's the same instinct behind how I've started using AI in this work: generating role-play scenarios and reinforcement material at a scale that lets every rep get specific, practiced coaching regularly, not just whenever a manager has a free hour. The tool changed, but the underlying belief didn't — consistency comes from structure, not from hoping people show up at their best every day.
The honest caveat
None of this means the people don't matter. A bad system executed by a great team still produces decent results, because people compensate. A great system executed by a disengaged team eventually breaks down too — systems reduce how much performance depends on individual mood or skill, but they don't replace the need for people who actually care about the outcome. The goal isn't to remove the human element from customer experience. It's to stop asking the human element to do all the work alone.