At 6:15pm, I receive a notification I’ve never seen before.
Not a complaint.
Not a churn alert.
Not an escalation.
A recommendation.
Retention action: None.
I open the account.
The customer has been with us for six years.
Regular purchases.
Moderate spend.
A few support interactions.
Nothing unusual.
Except one thing.
The AI predicts they’re about to leave.
Ninety-two percent confidence.
The dashboard highlights the conclusion.
Customer lifetime value remaining: Negative.
I stare at it.
Because that’s not how we used to think about customers.
Not people.
Not relationships.
Not loyalty.
Value remaining.
A retention specialist joins the call.
She looks uncomfortable.
“The system is recommending we let this one go.”
Let them go.
Like we’re discussing inventory.
Not a customer.
The AI explains its reasoning.
Acquisition cost recovery: Complete.
Future revenue probability: Declining.
Support cost trend: Increasing.
Predicted profitability: Negative.
The recommendation appears again.
Do not intervene.
At 6:24pm, I ask for alternative scenarios.
The model generates them instantly.
Offer a discount?
Negative ROI.
Offer loyalty credits?
Negative ROI.
Escalate to a specialist?
Negative ROI.
Every path leads to the same answer.
The customer costs more than they’re expected to contribute.
The machine has made its decision.
What unsettles me isn’t the calculation.
It’s the certainty.
For years, customer experience was built around retention.
Keep customers.
Build trust.
Earn loyalty.
Now the objective has changed.
Retain the right customers.
And quietly release the rest.
At 6:37pm, another recommendation appears.
Then another.
Then another.
The list grows.
Not one customer.
Thousands.
Each ranked.
Each scored.
Each assigned a future value.
Each evaluated against the same question:
Are they worth saving?
The dashboard calls them:
Retention Exceptions.
I call them customers.
The difference feels important.
Someone from finance joins the discussion.
“We’re spending too much protecting unprofitable relationships.”
Relationships.
The word hangs in the air.
Because relationships were never supposed to be optimised this way.
Not completely.
Not mathematically.
I click into another account.
The customer recently lost their job.
Order volume dropped.
Returns increased.
Support contacts rose.
The system interprets all of it as declining value.
I see something different.
A customer going through a difficult period.
The machine sees risk.
I see context.
The machine sees probability.
I see a person.
At 6:52pm, the final recommendation arrives.
Projected annual benefit of selective churn strategy: +$48.2 million.
The number is impressive.
The logic is flawless.
The shareholders would probably love it.
And yet something feels broken.
Because customer experience was never supposed to be about identifying who deserves help.
It was supposed to be about helping them.
I close the dashboard.
Outside, the city lights are starting to appear.
The office is getting quieter.
But the question stays with me.
If AI can tell us exactly which customers are worth keeping…
How long before we stop trying to earn loyalty at all?
Reality Check: How Close Are We?
Many of the concepts in this story already exist today:
- Predictive churn modelling
- Customer lifetime value forecasting
- AI-driven retention prioritisation
- Profitability-based customer segmentation
Businesses have always prioritised resources.
AI simply allows them to do it faster, more accurately, and at much greater scale.
The challenge is deciding where optimisation should stop.
CX Leader Takeaway
AI can help identify which customers are most likely to leave.
It can even recommend where retention investment should be focused.
But customer experience becomes dangerous when customers are reduced to financial forecasts.
The future of CX won’t be defined by which customers companies save.
It will be defined by how they treat the ones they choose not to.
Previous chapter:
Future of CX: Part 6 – 4:47 PM — The Perfect Apology
Next chapter:
Future of CX: Part 8 – 7:40 PM — The Customer You Never Spoke To
New Series: Future of CX
This story is part of a CX Today series following a single day in the life of a CX leader navigating AI, automation, and the growing tension between efficiency and humanity.
Each chapter explores what happens when customer experience becomes increasingly optimised by systems designed to maximise outcomes.
New chapter every week — next up: an AI agent resolves a customer problem before the customer even knows it exists.
For early previews and what’s coming next, follow Rob on LinkedIn.
