“What’s the future of the contact center in an AI-powered world?”
This was the question that Marc Bernstein, CEO of Balto, a real-time guidance platform, posed to a crowd during a recent contact center conference. The following presentation explored the growing capabilities of artificial intelligence, the increasing gap between what’s possible with AI and what’s socially accepted, and the reality that there will always be situations where people want to talk to people.
It was that last part, about people wanting to talk to people, where Marc spent most of his time. He shared a perspective on the need for capability and humanness in CX and challenged attendees to lean into their humanness.
It was a refreshing narrative, and I was interested in hearing more of Marc’s thoughts. We sat down for a discussion recently, where he expanded on his beliefs for how businesses can make the most of humans plus AI.
Here’s what I learned:
Justin Robbins: Marc, I was fascinated by what you had to say about Humanness and Capability during your presentation at the Call and Contact Center Expo. I appreciate you expanding on those thoughts with me today. Before we jump into that, though, who are you if people aren’t familiar with Balto? What do you do?
Marc Bernstein: I started Balto in January 2017. I’m the founder and CEO, one of three founders. I started Balto because I had a problem while on the phone myself. I was in sales, and my manager would give me excellent coaching advice, but I noticed that I would leave his office from coaching, get back to my calls, and apply nothing I had just been coached on. It was really painful.
I said to myself, how can I solve this problem of bridging the stuff that, you know, my management wants me to do, that the company has established as best practice, that I want to do myself, and then actually doing it in the moments when I need to. So in January 2017, we launched Balto. Balto offers real-time guidance, which is now commonly known as agent assist. We capture everything the agent says as they say it and everything the customer says as they say it, and give the agent live recommendations for them to be as effective as possible.
JR: Awesome, I love it. And as someone who spent many years delivering training and quality assurance for teams, I can vouch that the reality you experienced while on the phone is the case for many others. Leaders will have these coaching sessions, and either nothing is taken away from it, or the agent just picks and chooses out of everything they were told; which one do I want to work on? I appreciate how you’re working to solve that problem.
All right, AI was a big part of the conversation at our conference, which included your session. Let’s start with this question: What is your hot take on the current state of AI and the contact center?
MB: My take is that we are just about to cross the line where AI is better than people at most of those tasks. And what that means is that we have to figure out, first of all, how we have the people keep up. If you have a fully automated experience driven by AI, where can you have AI plus person be stronger than AI by itself? That’s the only way we’re going to have people keep up.
If you have just the human versus just the AI, think about it. An AI can process something like a hundred million bits of information per second. Humans have a thought, like a real tangible thought, every two to three seconds, right? And you know, think about how much time it takes to form a sentence. AI has done a hundred million calculations in that bit of time. And then AI has infinite memory and infinite storage. And by memory, I meant the ability to remember things and store all this information and retrieve it whenever it needs it.
Try to remember every single detail of your life, every single day, every word you read in every book. That’s what AI today can do.
So, as AI progresses, how soon can it take all of this data and all this information and apply it in a way where it’s coming up with better logical outcomes than a human could? AI is getting very close. We’re going to see what’s left for the human roles is people augmented with AI, kind of taking that super-powered memory and taking that ability to retrieve information and trying to augment people’s very fable memories but great ability to connect with other people and incredible ability to be creative and problem solve.
JR: You said we’re getting close. What does that timeline look like for you? Is that the next 12 months? Is it the next 12 years?
MB: Two to three years. If you had asked me six months ago, I would have said, you know, we’re doing some good stuff dabbling with AI; there’s some cool generative AI out there, it’s kind of creating content and blog posts, but this was before we saw ChatGPT, GPT 3.5, and GPT 4 come out, Google introduced Bard, etc. This powder keg kicked off six months ago, and now we’re in the midst of an explosion. For example, we’re almost at the point where AI can write all your blog posts and write them better than a human. In fact, in a lot of cases, it already can. So I thought it would have been five-plus years, but the last six months gave me a little bit of a shock and said, whoa, I think we’re looking at two to three years.
JR: You say that, yet I think about what must happen in two to three years to realize that. When I think about any innovation, the gap between what could be done and what’s adopted and being done can be vast. There’s getting to the point of capability on the technology side, but then there’s getting organizations to adopt and leverage it effectively. What do you see as some of the barriers to the adoption of AI in the contact center? What must happen to get us to the future state you’re predicting?
MB: Justin, that’s a critical point. I think we’re going to see the adoption of AI almost feel like a brand new teenage driver, where they’re pushing on the gas and hitting the brake and pushing on the gas and hitting the brake because they’re still trying to figure out how do you make the car run smoothly.
What we’re going to see is two to three years, the technology will be there, and everyone’s going to look at it and go, yup, this is capable of doing a lot of functions in the contact center. And then businesses themselves will have to grapple with how we implement this.
Businesses must consider things like, “How do we ensure that the AI doesn’t make mistakes?” It inevitably will, and when it does, how do businesses handle that? How do they handle it responsibly? How do they take the legal liability of the AI making a mistake with our customers? Can they prevent the AI from providing recommendations unless it has high confidence? Those are the sort of practical questions that will have to be worked out.
In addition to compliance and politics, we will see AI take jobs and create new ones. We must look at that seriously and soberly and discuss that realistically as an industry. It’s not just going to affect the contact center. We’re going to see jobs being taken everywhere, in every sector. When that happens, we’re going to have to continue to navigate that conversation in the political space, in the social space, and figure out how to protect folks who perhaps are more vulnerable and haven’t had the opportunities that other people have and make sure that there are still ways that those individuals can get livelihoods.
AI is going to not too soon, or not too far off, be so good that it’s not even like if you go to school, get a four-year degree, and become an expert in a space that you’re safe. AI is going to be incredibly, incredibly capable.
We need to look for the parts of experiences we have with companies and experiences we have day to day to learn what we value because they are human. Things we value because it is a person connecting with a person when human value is inherent. Those are the sort of experiences that will remain with people and not AI.
We will continue to try to separate the stuff that we can automate from human stuff, and that process will take time and that process will have some friction.
JR: I was very interested in learning more about this part of your presentation – this lens of looking at capability and humanness. If I lead a customer experience team, and I have to think through this idea of capability and humanness, what does that mean? How do I take that kind of approach? Can you share more of your thoughts here?
MB: Historically, the perspective was we’re going to apply people where people are clearly better than machines. And by better, I mean the machine can’t solve the problem. The machine can’t retrieve the necessary information. The machine can’t understand the context or the nuance behind the customer’s need. And we’ve said, “Well, in the places where AI can’t do that stuff, that’s where we’ll have people.” That paradigm is about to break because AI will be able to do all of it.
In terms of capabilities, we have to start assuming there are two options. You have an AI that can do the job, or you can have a human that can do the job, and they can do it equally as well. The question is, where does it benefit to have a human doing it? Not just complete the task, but where can the human give something a little bit extra, a little bit special?
For example, one of the things that folks often talk about is a human being able to comfort another human. I imagine that an AI could give you comforting words, and an AI could give you a message that you want to hear. But I think that just like how getting comforted by a dog has this very beautiful, intimate kind of relationship where you and the dog are just feeling this sort of connection; I think that humans will have a leg up on AI and having that sort of connection.
So, if you have a truly difficult or sad or frustrating scenario, humans will have a leg up on AI and having that sort of connection. It might feel like you’re getting pushed through a funnel if you get sent to an AI, even if the AI can handle your problem. Whereas at that moment, you need a person who can handle the problem just as well but can inject that humanness into it.
I’d say continue to look for areas where your customer experience needs humanness and start building a human brand now. That’s what some of the best leaders in the contact space have done. That’s what they are doing. I think that that’s what every contact leader should be thinking about right now.
JR: I love the intention behind this and am aligned with you that there will be certain instances where the connection of another human is critical. That said, I still see a lot of grey areas that could be open for interpretation and subjectivity on what elements of humanness really do matter coming from a human. You alluded to this a bit, and I’ve certainly had people try to convince me of some extreme cases, but how do you see AI replicating humanness?
MB: Consider natural language understanding (NLU). Software engineers have been communicating with computers using computer code for decades, so why does anyone want computers to understand natural language? Natural language is universal – no esoteric programming language is required, so everyone can interact – and natural language is better aligned with how we naturally think. People think in concepts, not computations. The NLU capabilities companies like Balto are building today are in response to the market wanting machines to operate in a more human-centric way. AI has been replicating humanness for decades. We just keep moving the bar of precisely how human we expect AI to be.
JR: Okay, fair enough. Let us consider the impact of humans on this. You encouraged contact center leaders to lean into their humanness during your presentation. What is your advice on practical ways to do that?
MB: Culture, context, and connection.
- Culture: Hire great people who want to grow with your company. Embed training and development into your contact center’s day-to-day. Set the expectation that taking great care of your customers is a serious and meaningful responsibility.
- Context: Don’t let policy be a substitution for rational thought. Develop and drill your plays, and also treat customers as individuals. Invest in clear, customer-centric processes for exceptions and escalations.
- Connection: Treat every customer interaction as a mini-advertisement for your brand with a captive and engaged audience. Consider the scale at which the contact center can shift customers’ attitudes and behaviors. Fanatically seek opportunities to increase customer loyalty, and act on them.
JR: Great advice, Marc, and we could spend a few conversations exploring each of those. Before I let you go, do you have any other advice for contact center leaders who are thinking about the role of AI and humans in service?
MB: Yeah, invest in AI automation and augmentation at the same time. Automation upstream, agent augmentation downstream.
If you invest only in automation, you’ll create fires for yourself. Your easy interactions will get deflected through self-service, and agents will see their jobs getting harder without additional resources to handle them. They will naturally begin to worry that the AI is coming for them next, which will strain morale and spike attrition.
Show agents you’re investing in them, too, by proactively driving agent efficiency through technology. Don’t wait until your contact center culture has gone to $#!% to introduce technologies that help the floor be more productive and better serve your customer.
JR: Thank you for sharing your thoughts with me today, Marc! I’m excited to watch how customer experiences evolve over the coming years. I hope organizations recognize how to make the most of their humans as AI becomes an increasing part of the equation.
Maximize the Impact of Customer Service
Join us on June 29th as Balto hosts Justin Robbins for a special presentation about How to Deliver Ridiculous Value with Your Customer Service Team.