The Hold Music Is Dead

Last month I called my doctor's office.

The first rep picked up and asked me to verify my date of birth. Then she transferred me. The phone rang. And rang. A second rep picked up and asked me to verify my date of birth. Then transferred me. The phone rang. And rang. A third rep picked up and asked me to verify my date of birth.

I hung up.

Three people. One office. Fourteen years as a patient. No one knew who I was.

In that moment, the office told me everything I needed to know about how they actually feel about me.

They don't.

Twenty minutes later, I opened a streaming app on my phone. Without me typing a single word, it suggested a documentary about college football. I love college football. Nobody at that company has ever met me. Nobody there knows my name. And yet they knew me better than the doctor I had been seeing for 14 years.

Same person. Same phone. Two completely different worlds.

One of those worlds is about to disappear. Most leaders have no idea it's already happening.

The 95% Problem

Here is a number that should keep every executive up at night.

Most companies use less than 5% of the customer data they already have. The other 95% just sits there. It sits in call recordings nobody listens to. It sits in chat logs nobody reads. It sits in survey responses nobody opens. It sits in the body language of a video call nobody watches a second time.

We don't have a data problem. We have a memory problem.

Imagine if every time you walked into your favorite coffee shop, the person behind the counter looked at you like a stranger. Even though you've been coming there every morning for three years. That's what most companies do to their best customers, every single day.

The wild part? They paid for the memory. They just never turned it on.

From Answering to Remembering

For 30 years, customer service has been built around one job: answer the call.

Pick up the phone. Read the script. Solve the ticket. Close the case. Next.

That model is dying. Not because anyone wants it to die. Because customers stopped accepting it. People who can order a sandwich with their thumb in 90 seconds will not get passed around three times just to confirm their date of birth.

The new job is different. The new job isn't to answer. It is to remember.

Recognize who's calling. Be ready for what they need. Pick up on how they're feeling. Not to be intrusive. To save them time.

That sounds like science fiction. It isn't. It's already shipping.

What AI Actually Does Here

Let me explain it the way I would explain it to my kids.

Picture a brand new employee on their first day. They are smart, but they don't know anything about your company. So you give them three gifts.

The first gift is a giant notebook with every conversation your company has ever had with every customer. They read it all in one afternoon.

The second gift is a pair of glasses that can tell, in real time, when a customer's voice is shaking, when their words are short, when they are about to give up. The glasses whisper, "She's frustrated. Slow down."

The third gift is a tiny voice in their ear that says, "The last three people who said exactly what she just said ended up canceling. Here are the two things that saved them."

That's it. That's what AI does in customer experience.

It doesn't replace the human. It makes a brand new employee act like a 20-year veteran on day one. And the volume tells you it's already happening at scale. Verizon prevented 100,000 customers from churning in 2024 by using AI to analyze 170 million annual calls. That's 170 million conversations analyzed by one company in one year.

A Tale of Two Calls 

Let me tell you about two hypothetical calls.

The first call. A woman named Linda calls her insurance company. Her husband just passed away. She needs to update her policy. She gets routed to a queue. She waits. She finally gets a rep who says, "Can I have your policy number?" Linda doesn't know it. She starts to cry. The rep gets flustered. The call ends without solving anything. Linda cancels her policy three days later.

The second call. A man named Marcus calls a different insurance company. His mother just passed away. Before the rep even picks up, the system has already connected the death certificate Marcus uploaded last week to his record. It has already flagged this call as sensitive. The rep's screen shows two words in plain English: "Be gentle." The rep picks up and says, "Marcus, I'm so sorry for your loss. I see what's going on. I'm going to walk you through this slowly. Take your time."

That's the same technology that handles your password reset. Used differently.

One company saw a ticket. The other saw a customer. Guess which one will still be in business in 10 years.

How Marcus' Call Actually Happens

I want to pull the curtain back for a minute, because the second call is not magic. It's a real architecture, shipping today, in companies you've heard of.

There's a platform called Genesys. They run the contact center for over 8,000 companies worldwide and roughly 2 million daily users. They are not a chatbot startup. They are the plumbing behind the calls you make to your bank, your airline, and your insurance company. And in the last two years, they have quietly rebuilt what a call is.

Here's how the second call actually works.

When Marcus dials in, the system isn't just routing him to the next available person. It is already running. A software solution called the Event Data Platform stitches together the things Marcus has already chosen to share, the form he started last week, the upload he made, the chat he completed. That data lives in one place, not 14 different systems. By the time the rep picks up, the platform has already connected the death certificate from last week's upload to Marcus's record.

Then a second tool kicks in. Genesys calls it Agent Copilot. Think of it as the human rep’s partner on her screen. It monitors the call in real time, notes the pauses, suggests the next sentence, pulls the right form, and drafts the wrap-up notes so the rep isn't typing while a grieving customer is talking. Companies using it report five minutes less handle time per call, 90 seconds less hold time, and two minutes shaved off after-call work. Multiply that by a million calls a year...that's a different business.

Then there's a layer called AI Studio. This is where the company builds its own AI helpers without writing code. A supervisor can type, "Build me an AI helper that flags any call where the customer mentions a death in the family, and tells the rep to slow down." That's it. That's the build process. What used to take a six-month IT project now takes an afternoon.

And here's the part that matters most. Every one of these pieces is governed. The AI doesn't run wild. It has guardrails. While it logs every action, the system asks the human for confirmation before it changes anything sensitive. Because the only thing scarier than an AI that doesn't help is an AI that does help unregulated.

This is what people mean when they say "agentic AI." It isn't a chatbot. It's a team of small, specialized AI helpers that watch the data, suggest the next move, and only act when a human says yes.

What Actually Changes When You Turn It On 

I'll give you a real example. Genesys ran its own product support team through this same stack. Their own technology on their own people. Here is what happened in three years.

A 20-point jump in customer experience score. 35% of calls shifted to chat without losing quality. Routing time cut by 34%. Three to five minutes saved per call. Case escalations down 43%. Employee attrition under 3%, in an industry where 30% is normal. And 157,000 working hours saved.

157,000 hours. That's 75 people working full-time for a year. Recovered. Without firing anyone. Without working anyone harder. Just by turning on the memory that was already there.

Further, a healthcare provider in Chile, RedSalud, ran the same playbook. 20% more patient bookings. 30% lower contact center costs. Higher satisfaction. That's not a tech story. That's a healthcare story. That's more people getting seen by a doctor because the phone system figured out who they were.

Three Shifts You Need To See

I want to give you the simplest possible map of what's happening. Three shifts...

Listening got cheap. It used to cost a fortune to figure out what customers were saying across phone calls, emails, chats, and reviews. Now the financial burden is significantly reduced. Every word, every pause, every sigh can be turned into a signal. AI can score every interaction automatically, in the supervisor's native language. A year ago, supervisors spot-checked maybe 2% of calls. The other 98% disappeared. Now the moments that matter, the call where a customer was struggling, the chat where a rep nailed it, the pattern that says "we have a problem here," none of that has to be lost. The companies winning right now decided to actually use those signals instead of letting them rot.

The rep stopped working alone. For decades, the person on the other end of your call was on an island. A screen full of windows, a script in their lap, a clock ticking down. Now they have a partner sitting next to them that never sleeps. The partner surveys every word, suggests the next sentence, pulls the right document, and intimates when to pause. The number of reps using these copilots tripled in a single year. Not in pilot. In production. Across thousands of companies.

The customer stopped being a stranger. This is the big one. Tools like Predictive Engagement note when a visitor is struggling, three failed searches, a form abandoned twice, a price page visited five times, and provides the rep a heads up before the customer ever says hello. Every time you call, chat, click, or complain, the company learns something. The companies turned that learning into ‘greet you like a regular at a coffee shop’.

The Line Between Helpful and Creepy

I know how this can sound. The more I describe what these tools do, the more it can feel like the company is always watching you. That's a fair worry. Worth saying out loud.

Here's the thing. There's a red line that matters. We all know it. We're customers too. The line is the difference between "I remembered you" and "I followed you."

A barista who remembers your order? Helpful. A barista who reads your texts? Creepy.

The companies that win earn trust with permission. They use data the customer chooses to share. They explain what they're doing with it. They make the experience feel like service, not surveillance. They stop asking questions the customer already answered. They stop making people repeat themselves. They show up prepared.

The point isn't to know everything about a customer. The point is to stop treating loyal customers like total strangers.

The Real Question Isn't Technology

Here's where I have to be honest with you.

Every company I work with thinks the question is, "What tool should we buy?" That's the wrong question.

The right question is, "What kind of company do we want to be?"

AI doesn't change who you are. It just makes you faster at being who you already are. If your culture treats customers like tickets, AI will help you treat them like tickets at scale. If your culture treats customers like people, AI will help you treat 10 million of them like people, all at once, every single day.

The technology is the cheap part. The culture is the hard part.

The companies you'll see win in the next five years aren't the ones with the biggest AI budget. They're the ones who decided, before they bought anything, they wanted every customer to feel known. The tech is just how they keep that promise at scale.

The hold music isn't dying because customers got impatient. It's dying because we finally have the tools to do better. The companies that figure this out won't just sell more. They will be remembered. Being remembered, in a world where everyone is competing for eight seconds of attention, is the only moat left.

So here's the question I'll leave you with.

The next time one of your customers hangs up the phone, what will they remember about you?

They will remember something.

The only question is whether you got to choose what it was.

 

Want to talk through how this applies to your environment? Reach our team at data.analytics@trace3.com.