Why Your Contact Center Still Needs Humans

Picture this: You’ve lost something important, you know exactly where you left it and you need someone to help you get it back. But instead of getting help, you get an automated email message from a contact center, over-and-over-and-over-and-over-and-over again.

That’s what happened to customer service expert, Jeff Toister, when he lost his jacket in a rental car in the Spring of 2024. At ICMI Contact Center Expo, he shared this story, which illustrates why contact center staff shouldn’t be so AI-focused.

First, he got a response on social media from an agent named “Blake.”

Hello. I'm sorry that you've not heard from our location in Newark yet. I am showing that the status is still pending. Please allow more time for them to find the items and contact you. -Blake

“Even when we put people in front of customers, they sometimes act like bots,” Toister says. “A message like this can no longer be acceptable in the age of automation. If I want to send a customer a message that says ‘keep waiting, we’re not trying,’ I can automate that.”

Then, over the course of a month, he kept getting this AI-generated message:

Unfortunately, your lost coat/jacket has not been located at this time. We will continue to search for your lost item.

“I don't know what all the stages of grief are for a lost jacket, but this message felt like an eye poke,” Toister says.

Toister isn’t anti-AI. He shared several use cases, from Fertitta Entertainment to reducing their abandonment rate from 40% to 10% by using an AI voice agent, to Allstate discovering that AI writes more empathetic, jargon-free emails for accident claims. The key is knowing when to use AI.

“Automate the CRaP (Confident, Routine, and Predictable) — the crap that your customer doesn’t want to wait to deal with a human for,” Toister says.

Once you’ve done that, here are a few moments where you need your humans to show up:

• Saying hello. When Toister observed more than 400 customer interactions in person, he found that when employees smiled during greetings, customers returned the smile almost 70% of the time. “A great hello makes people feel welcome,” Toister says. “It's not the script that marketing gave you, it’s not saying the right words, it’s making your customers feel welcome.”
• The point of engagement. Giving customers the option to reach a human — even when most won’t use it — improves satisfaction and conversion. Studies show that anxious customers who couldn’t access humans were the least satisfied, yet only used the human option 16% of the time when it was available.
• Listening. Using AI to analyze over 29,000 calls, Toister says that agents who made customers happy asked nearly five more questions per call. Contrary to conventional wisdom, these empathetic agents weren’t less efficient, they were more so.t less efficient, they were more so.
• Being helpful. Toister shared the story of a woman whose daughter’s lacrosse bag went missing and who could track it to the exact baggage office in Chicago via an AirTag, only to be told by multiple agents, “I'm not allowed to contact Chicago.” So, she flew to Chicago to pick it up herself. Yikes. This story made national news, BTW.

The Business Case for Agents

Headset Advisor ran an experiment comparing AI bots to human agents for pre-sales questions. The AI achieved an 18% satisfaction rate. Human agents? 92% satisfaction, with twice the conversion rate and an average order value of $300.

Love’s Travel Stops implemented an AI voice agent for roadside assistance and reduced average speed of answer from nine minutes to nine seconds, but they also gave customers the option to wait for a human. The result? Agent turnover decreased by 50% because the humans who did take calls were dealing with customers who genuinely needed them, not truckers who’d been on hold for nine minutes.

At Amazon’s distribution centers, every package gets weighed before shipping, and if the weight doesn’t match expectations, humans are alerted to look at it. Built into the system is quality control: automation handles the routine, humans handle the exceptions.

“We need to enable our agents to take those steps to do the right thing and we need to be okay with agents just doing it and asking for forgiveness later," Toister says. “That’s a humanity piece that’s often missing where we hide behind rules and where we make things unnecessarily complicated.”