CX has spent the more than 20 years building programs around a simple idea: listen to customers, respond when something goes wrong, and fix the bigger issues when patterns appear.
That idea helped the function mature. It gave teams language, process, governance, and legitimacy at a time when most companies treated customer feedback as interesting but optional.
What Bain called the Closed Loop Feedback (CLF) model, consisted of inner-loop and outer-loop framing; and it had real value. The inner loop created accountability for individual customer recovery. The outer loop gave companies a way to look across feedback, find root causes, and improve the business. For many organizations, that was a major step forward. It moved customer listening out of research folders and into operating routines.
But every framework has a shelf life.
The problem now isn't that closed loop feedback was wrong. The problem is that too many CX teams are still treating it like the full answer when it isn't. Customer signals are no longer limited to survey responses and relationship feedback. They live in service calls, chats, emails, usage data, complaints, reviews, renewal patterns, sales notes, product telemetry, digital behavior, and social conversations.
Customers are telling companies what's working and what's broken every day, often without filling out a survey at all. The business environment has changed too much; most notably in the last three years with the advent of Large Language Models (LLMs) and enterprise focus on artificial intelligence.
That's why CX needs to evolve to a four-loop operating model.
What is the four-loop operating model?
The old model taught companies how to respond and improve. The four-loop model gives CX a way to recover, remove, orchestrate, and learn.
It's a nuance perhaps, but the nuance matters because the pressure on CX is very different now. Executives aren't just asking whether customers feel heard. They're asking what the company changed, what revenue was protected, what cost was removed, what risk was reduced, and how the CX function is impacting the company strategy.
If CX keeps returning to old frameworks, its influence will keep shrinking. Budgets will follow, and we've already seen this trend accelerating.
This is happening because CX hasn't consistently translated that understanding into business value with enough discipline.
So I've evolved the CLF model to accommodate today's reality and help CX teams evolve from the dogma we've been trained to accept. The remainder of this article walks you through each of the loops, and helps CX leaders better prepare for the moment we're in right now.
Loop 1: Recovery
The first loop is recovery. This is the part of closed loop feedback most teams know well. A customer has a problem, the company responds, and someone tries to make it right. That still matters. When a customer is frustrated, ignored, confused, or angry, the response they get can either preserve the relationship or confirm that the company doesn't care.
Recovery is human. It's also necessary. But recovery can't be the high-water mark for CX.
Too many programs have confused case closure with customer value. A customer gets a call. A ticket gets closed. A dashboard turns green. Internally, everyone can point to activity. Externally, the same issue often shows up again for the next customer, and the next one, and the next one after that.
That's where the second loop matters.
Loop 2: Removal
Removal is the discipline of taking defects out of the business. It asks a sharper question than traditional outer-loop work usually asks. It doesn't stop at "What themes are we seeing?" It asks, "What exactly is broken, who owns it, when will it be fixed, and how will we know the pain is gone?"
This is where I think CX has been too polite.
We've allowed too many recurring customer issues to be treated as insights instead of defects.
An insight can sit in a deck. A defect needs an owner. An insight can be discussed in a quarterly readout. A defect needs a deadline, a fix, and proof that the problem stopped repeating.
That shift changes the role of CX. The function shouldn't be the team that merely describes customer pain. It should help the company prioritize which pain matters most because of the value attached to it. Which issues put renewal revenue at risk? Which ones increase service cost? Which ones slow adoption? Which ones create repeat contacts? Which ones damage trust with the customers the company can least afford to lose?
This is how CX earns a more serious seat at the table. Not by asking leaders to care more about scores, but by showing how customer friction turns into financial drag.
Loop 3: Orchestration
The third loop is orchestration, and this is where the AI conversation creates a real opening for CX.
Every executive team is trying to figure out where AI belongs. Most companies have started with obvious productivity use cases. Summarize calls. Draft emails. Search knowledge bases. Automate tasks. Reduce handle time. Those use cases are useful, but they don't reach the bigger opportunity.
The bigger opportunity is using AI to help the company decide what to do next for customers.
That's where CX teams have a wedge. They're sitting on years of customer insight that most companies have underused. Survey comments, complaint patterns, service reasons, relationship feedback, research transcripts, closed loop notes, frontline observations, product friction, churn themes, onboarding pain, and renewal risks all have value far beyond reporting.
For years, that corpus has mostly been used to explain what happened. Now it can help drive what should happen next.
A customer shows early signs of renewal risk. What should the account team do?
A product user keeps hitting the same friction point. What should the company change in the interface, onboarding, or support content?
A high-value customer reports a repeated billing problem. Who needs to intervene, what message should they use, and what policy or process should be reviewed?
A segment of customers is quiet in surveys but declining in usage. What action should trigger before the relationship weakens?
This is orchestration. It turns customer signals into guided action.
CX needs to be in the middle of that work. Not because CX should own every action, and not because CX should pretend to be the AI engineering team. CX should be there because the function owns the deepest understanding of what customers are telling the company, where the pain is recurring, what moments carry the most value, and which actions are likely to matter.
That's a more powerful role than managing a survey program.
It also challenges one of the oldest assumptions in the function: that listening is the core job. Listening is only useful when it changes what the company does. In an AI-enabled operating environment, listening becomes the raw material for action design.
The customer signal has to feed rules, recommendations, workflows, playbooks, and models that help employees make better decisions in real time.
This is where closed loop feedback starts to look too small for the moment we're in.
The inner loop helps with recovery after a customer raises their hand. The outer loop helps with improvement after enough patterns are visible. But modern companies need to sense, decide, and act much faster than that. They need to respond to direct feedback, indirect behavior, emerging risk, and value signals at the same time. They need customer understanding to move into the systems where work actually happens.
That leads to the fourth loop
Loop 4: Learning
Learning is the discipline of writing outcomes back into the system so the company gets smarter over time. This is the loop most companies talk about and rarely build.
A team takes action. Did it work? Did the customer stay? Did usage recover? Did repeat contact decline? Did the product fix reduce confusion? Did the save motion protect revenue? Did the onboarding change improve adoption? Did the AI recommendation lead to the right employee behavior? Did the playbook actually help, or did it just create another internal step?
Without the learning loop, companies stay busy but don't compound knowledge. They keep acting on the same problems with slightly different language. They keep treating customer feedback as a source of stories rather than a source of operating intelligence.
The learning loop forces a more honest standard. It says the company shouldn't just close cases, launch fixes, or trigger actions. It should measure whether those actions created value and then use that evidence to improve the next decision.
Collaboration and ownership in the four loop model
This is also where finance needs to be part of the model. CX has spent too much time trying to defend itself with score movement, correlation studies, and broad loyalty claims. Some of that work has merit, but it often feels too far away from the business decisions executives are making.
The four-loop model creates a clearer value language.
- Value at risk: where customer pain threatens revenue, margin, adoption, growth, or trust.
- Value removed: where defects have been taken out of the business and no longer create avoidable cost or friction.
- Value created: where better actions drive retention, expansion, conversion, usage, or lower service burden.
- Value learned: where the company improves its models, playbooks, policies, and decisions because it measured what worked.
That language is harder to dismiss than another debate about NPS movement.
It also creates a different operating rhythm.
Recovery can happen continuously.
Removal can run through bi-weekly value standups where CX, product, engineering, operations, and service agree on priority defects.
Orchestration can connect customer signals to sales, marketing, service, and customer success actions.
Learning can be reviewed monthly with finance and executives so the company can see which actions protected value, created value, or failed to produce the expected result.
This is the kind of model CX needs because it makes the function harder to marginalize. It shows where customer understanding changes the business, not just where customer feedback is collected and reported.
There's an important leadership point here, too. CX teams shouldn't try to own all four loops by themselves. That would be a mistake. Recovery belongs close to service and customer success. Removal has to involve product, engineering, operations, and policy owners. Orchestration needs sales, marketing, service, customer success, and data working together. Learning requires finance, analytics, and executive review.
CX should set the priority and the value lens. That's the job.
The function should make sure customer signals aren't softened into vague themes, buried in dashboards, or converted into internal activities that don't change anything, or don't change anything meaningful. CX should be the group that keeps asking whether the company is making the right customer issue visible, assigning the right owner, triggering the right action, and measuring the right outcome. CX needs to be pushing their partner's thinking, not tracking their activities. That role is far more relevant to the next decade than running a traditional closed loop program.
The dogma that needs to be challenged is the belief that closing the loop is the proof of customer commitment. It isn't. All it does is tell the customer, "We heard you." That's good, and all the big CXM vendors will tell you that follow-up increases NPS, but it still doesn't prove the company changed at all. It doesn't prove the problem went away. It doesn't prove the next customer will have a better interaction. It doesn't prove the business learned anything.
CX has to stop treating response as the end of the work. We've been doing that for far too long. And I'm not surprised given I get way more questions about how to increase responses rates than I do about how to increase the velocity of action.
The standard must be compounding value. A signal comes in. The customer is helped when help is needed. The defect is removed when the business is creating repeat pain. The right action is triggered when value is at stake. The outcome is written back so the next decision gets better. And while easier said than done, that's the model.
It's also the clearest path for CX to enter the AI conversation with credibility. Instead of talking generally about personalization, prediction, journeys, or VoC, CX can bring a specific asset: the company's customer insight corpus. It can show where that data can improve next-best actions, risk detection, frontline guidance, product decisions, and learning systems. Most companies have decades worth of feedback.
It's a serious wedge. So, while most functions are trying to figure out what data they have that can make AI more useful, CX already has data that explains customer pain, behavior, expectations, confusion, trust, loyalty, and risk. The issue is that many CX teams still treat that data as material for reporting on NPS trends instead of fuel for action.
That has to change. And my hope the change starts after you read this article.
Building measurable influence and delivering impact
If CX keeps leaning on 25-year-old frameworks as its primary operating model, the function will keep getting judged by 25-year-old expectations. Survey response rates. Closed cases. Score movement. Readout quality. Executive visibility. Those things won't be enough to protect influence in a business environment that's becoming more automated, more financially disciplined, and less patient with functions that can't prove impact.
CX doesn't have to abandon closed loop feedback, much like it doesn't have to abandon NPS. It just needs to put it in its proper place. Recovery remains essential. Root cause work remains essential. But the model has to expand because the work has expanded. Customer signals now need to drive action across the company and learning back into the system.
That's the shift from a feedback program to an operating model, and it's the shift CX needs to make now.
The future relevance of customer experience work won't come from better listening alone. It must come from helping the company turn customer signals into decisions, actions, fixes, and learning that can be measured in business terms.
Closed loop feedback gave CX its first real operating language. The four-loop model gives CX a way to matter again.
Quick links
- Where the AI vendor market is going, and what it means for CXOs
- A role designed for CX in the age of AI
- Intent drift is the new CX measurement gap (and your dashboard can't see it)