AI: Amazing, Impressive… and Absolutely Not Ready to Answer Your Customer’s Call

By Marc Bombenon

I’ve been in the call center business for over 3 decades, and I’ve watched technology move from rotary phones to omnichannel AI dashboards that claim to know what your customer meant to say before they finish saying it.  Impressive? Absolutely. Ready to replace your live agents? Not even close.

Let me explain before someone in Silicon Valley sends me a holographic fruit basket.

The Bleeding Edge, Emphasis on Bleeding

Engineers are doing amazing work in labs. I’ve seen AI answer questions with charm, humor, and frightening precision in perfect conditions with flawless audio, clean data, and no background noise.   Then they package it out to the real world.

That is when the bleeding starts.  Once you get past the “cool factor,” you start to realize it doesn’t always work, and always a huge subjective, meaning sometimes it doesn’t work, like at all.  (See “Steve” below)

These systems are being rushed into production to secure more funding to fix the problems that should have been solved in the first place.

It is a classic case of “beta test on the public, bill it as innovation.”

Let Technology Do the Heavy Lifting

Manual rebate processing belongs in a museum. It’s slow, error-prone, and expensive. Automation changes the game submissions, validations, and redemptions happen seamlessly in one system, reducing human error and cutting administrative costs by 30–50%.

Automation also opens the door to real-time data, your biggest value.   You can instantly see which offers perform, where bottlenecks exist, and how customers engage. That feedback loop lets you pivot quickly, tighten costs, and make data-driven decisions that directly protect your profit margins.

My Complicated Relationship With Siri

Now, I love Apple. Big fan. Loyal customer. Siri and I have been together since she could barely tell me the time.  So cool, she could tell me the time on the phone I was holding in my hand with the time displayed.  Whatever, still cool!!  Over time (years), she has grown, evolved, and even picked up a lovely female Australian accent.

But like a teenage daughter, some days she gets me, other days not so much.

This morning, I said, “Hey Siri, call Katelyn.”

And Siri replied, “Did you mean Steve?”   Really??

Now imagine that same level of “AI confidence” handling your best customers, your first-time caller, or your most frustrated client. 

That is not just bad, that is brand damage in surround sound.

When Math Meets Mayhem

AI does not actually hear your voice. your voice is mechanically captured, electrically transduced, digitized, and then interpreted by AI, which formulates and produces a digital response that is then converted back into a synthetic voice response that is mechanically reproduced. 

Every call goes through tens of thousands of variables such as accents, background noise, speech rate, emotion, and tone.

All it takes is one hiccup, and suddenly your AI has turned “I want to upgrade” into “cancel my account.”

That is not machine learning; that is customer churning.

Corporate Hall of Shame

Just last week, I called FedEx with a simple inquiry. No humans, only Agentic AI.

No comprehension. No escalation.

The digital voice repeated the same useless line again and again. There was no path to a human, and that was by design. The agents have been “disbanded.”

I was continuously urged to the self-service website, like I had missed that option the previous 4 times.  The same website that could not help me in the first place.

After five failed attempts, the AI hit the programmed call time limit and hung up, cleanly, efficiently, and without care or remorse.

I tried again and got the same result.  Right then, FedEx lost me as a customer, forever.

Using AI to save money by laying off people, providing poor service, and losing clients is a business model only a spreadsheet could love.

Congratulations, FedEx. Your AI saved four minutes of agent time and lost a lifetime customer.  About four dollars saved versus ten to fifty thousand lost.   Oops.

The Bottom Line

AI is here to stay, and it should be.

But AI is the perfect tool to enhance quality and efficiency for the humans handling your humans, not a replacement.  Otherwise, those humans that we call customers will replace you.

Empathy, understanding, and humor do not come from code; they come from connection.

Your customers want help, not hope that the algorithm gets it right this time.

Until AI can handle frustration, sarcasm, and human emotion without confusing my friend for Steve, it is not ready to take your calls.

Scalable. Secure. Smart. Onshore. ™

SureCall Contact Centers — where real humans still answer the call better than AI does.