I woke up at 6:30am.
Overnight, our AI resolved 12,842 customer issues.
By the time I opened my laptop, the contact centre had already done a full day’s work.
No queues.
No escalations.
No human interaction.
Just resolution.
At 9:05am, the message came through.
From the CEO.
“Customer support automation is at 94%.
Let’s get to 100% before Q3 earnings.”
I read it twice.
Not because it was surprising.
Because it was inevitable.
We’ve been moving in this direction for years.
First, it was chatbots handling simple queries.
Then AI agents taking full conversations.
Now, entire customer journeys—resolved without anyone ever speaking to a human.
The data is undeniable:
- Faster response times
- Lower operational costs
- Higher consistency
- Scalable to millions
The board sees efficiency.
The market sees margin.
The CEO sees share price.
But I see something else.
I open the overnight exceptions log.
There’s only one case left unresolved.
One.
Out of 12,842.
I click into it.
A customer is trying to cancel an account.
The request is simple.
But the context isn’t.
Their partner has passed away.
The AI followed protocol exactly as designed:
- Verified identity
- Requested documentation
- Provided next steps
- Maintained a polite, neutral tone
At no point did it fail.
At no point did it deviate.
At no point did it understand.
The customer replied three times.
Each message shorter than the last.
Until eventually:
“I don’t have the energy for this.”
The case is still open.
Not because the system couldn’t resolve it.
Because the customer gave up.
At 9:17am, I’m in a meeting with product and engineering.
They’re excited.
“Once we remove the final escalation layer,” one of them says, “we eliminate variability entirely.”
Variability.
That’s what we’ve started calling it.
The human part.
I ask a simple question.
“What happens in cases like this?”
There’s a pause.
Then a response:
“The system handled it correctly.”
Technically, they’re right.
At 9:32am, I draft a reply to the CEO.
I delete it twice.
Start again.
Because this is the decision.
Not just about cost.
Not just about efficiency.
But about what customer experience actually means.
We can get to 100% automation.
The technology is ready.
The data supports it.
The market will reward it.
But there’s something the dashboards don’t show.
Something the models can’t measure.
Something that doesn’t scale.
The moment a customer doesn’t need a resolution.
They need a human.
I hover over the send button.
And for the first time in years, I hesitate.
Reality Check: How Close Are We?
Many of the technologies in this story already exist today:
- AI-powered customer service agents resolving end-to-end interactions
- Emotion and sentiment detection in voice and chat
- Automated workflows for account changes, refunds, and cancellations
- Predictive CX systems reducing the need for human intervention
While full automation isn’t universal yet, the trajectory is clear:
Fewer humans. More AI. Greater efficiency.
CX Leader Takeaway
Automation will continue to redefine customer experience.
But efficiency alone does not build trust.
The real challenge for CX leaders isn’t whether AI can replace humans.
It’s deciding:
Where it shouldn’t.
Next chapter coming soon: 10:12 AM — The Empathy Algorithm
New Series: Future of CX — Humanity Against the Machine
This story is part of a new CX Today series, Future of CX: Humanity Against the Machine — a first-person walk-through of what it really feels like to run customer experience when AI agents, automation, and predictive systems begin making decisions before customers even ask.
Each chapter follows a single day in the life of a CX leader balancing human empathy with relentless pressure to automate, optimise, and increase shareholder value.
New chapter every week — next up: emotion detection, adaptive persuasion, and the line between helping customers and exploiting them.
For early previews and what’s coming next, follow me on LinkedIn.
