ujet.cx

“Human agents are the new CX differentiator.” Nah, humans have ALWAYS been the key. In a world where AI self-service experiences are getting worse before they get better, the organizations keeping humans in high-judgment roles are going to pull ahead on customer retention. Clock it.

Average contact center handle time is 648 seconds. That's 10.8 minutes per interaction. Deflection moves that cost from the humans your human customers actually want to speak to, to AI credit costs for a sh****y 'agentic' bot your customers hate. Contact elimination removes the cost all together. ujet.cx/spiral

Deflection is what you do when you don't know why customers are calling. Elimination is what you do when you do ...and you fix it upstream.

The contact center industry got very good at measuring containment and called it success. The market is learning the difference between containing failure and eliminating it.

The contact center isn't a cost center. It's where customer relationships are won or lost. Act accordingly.

Honest question: when did "AI in the contact center" start meaning "fewer humans" instead of "better humans"?

Eliminating one legacy ticketing and case management system saves $3M+/year. The stack was always the budget problem. Not the people.

Here's what a decade of watching AI "fix" the contact center taught us: 🧵

Customer service agents have no feedback system btw. You're burning money on preventable problems your frontline comes across DAILY. Hear what they have to say on Reddit (warning TRUTH ahead) reddit.com/r/callcentres/…

When was the last time your contact center asked not "how many did we deflect?" but "what would make this contact unnecessary?"

Every deflection metric you celebrate is a root cause you've decided to keep. No no, read that again.

Because the best interaction is the one that never has to occur. (because you fixed the issue before they had to contact you.)

"But how do I fix the issue BEFORE I know it's happening??!" - I hear you protest. Well... Spiral does this with an auto-generated issue taxonomy based off 100% of your conversational data. Alerting to you to those unknown issues hiding in sparse conversations BEFORE they become a fire to put out.

The 'AI success metric' contact centers have been tracking for a years is wrong. Not slightly miscalibrated, like genuinely structurally broken. 🧵 A thread on why your deflection rate is measuring failure and why your conversational data is the REAL unlock to AI success: