The Catalyst by Softchoice

How conversational AI is changing the customer experience

• Softchoice • Season 6 • Episode 18

Share episode

The future of business is a conversation. In this episode of The Catalyst by Softchoice, Heather Haskin and Craig McQueen, Vice President of AI Solutions at Softchoice, dive into how conversational AI is reshaping industries and redefining success.  

From smarter customer support to groundbreaking innovations, discover why this technology is more than just talk—it’s a game-changer. With real-world examples and insights into overcoming challenges, this episode is your ultimate guide to navigating the AI revolution.  

Featuring: Craig McQueen, Softchoice Vice President of AI Solutions 

The Catalyst by Softchoice is the podcast dedicated to exploring the intersection of humans and technology.

Heather:

You're listening to The Catalyst by Soft Choice, a podcast about unleashing the full potential in people and technology. I'm your host, Heather Haskin. The idea of talking to bots was once the imagined future of science fiction. Oh, how quickly things have changed. In just the last few years, conversational AI tools like ChatGPT, Copilot, Amazon Q, and Gemini have gone from experimental gadgets to essential tech for businesses. From virtual assistants and advanced chatbots to powerful automation tools, conversational AI is shaking up industries, redefining how companies operate, engaging with customers and competing in a digital first world. This technology isn't just an upgrade. It's a historic disruption. Today I'm joined again by one of my favorite guests we've ever had on the show, Craig McQueen, SoftChoice's Vice President of AI Solutions. We're going to discuss how soft choice customers and industry leaders are using this trailblazing tech and how conversational AI is rewriting the rules of business. Craig, welcome back to The Catalyst.

Craig:

Glad to be with you today, Heather.

Heather:

So you've been on the show several times. I know that our listeners have a pretty great idea of what you do, but today we're talking about conversational AI. So I'd love to hear from you what you do with this technology at SoftChoice.

Craig:

Sure, I'd be happy to, Heather. The responsibility of myself and team is really to help provide some clarity with customers on how AI can help their business, and then build systems that they can implement so they can take advantage of that technology. And certainly there's been a lot of headlines, a lot of noise and confusion around AI. And so we've had more customer conversations, helping them understand implications for the business this year than I've ever seen in my career in technology services.

Heather:

And with all those customer conversations, who are you interviewing? Who are you talking to? Who's asking you about conversational AI?

Craig:

A great question. We started this year speaking with IT leaders around generative AI. And guess what? Other people started showing up. So I would be in meetings, and the VP of Marketing would show up. Or the Chief Operating Officer. There's tremendous interest from line of business on what does this really mean, and they want to understand it. Which is different historically. We haven't had this in a long time. You might have had it 30 years ago when a VP of Marketing might want to know well, how should the internet affect my marketing plan? And so this is another evolution of A general purpose technology that impacts all lines of business. And so the heads of the lines of business are trying to understand what impact and benefit could I get. By introducing AI into what my organization does.

Heather:

So I love what you just said there. You said that organizations are asking, what can this do for my organization? And that's an interesting way to look at a problem. You want to solve a problem, but it sounds like they don't know what that problem is. So I would ask you, Craig, how do you typically come into a conversation with organizations and have that art of the possible conversation? Who is involved in that? Who are the stakeholders and how do you typically solve that for them?

Craig:

Sure. With any transformational change, you need strong executive commitment. So it wouldn't make sense to have any AI initiative unless you have an executive putting up their hand and saying, I will sponsor this within my organization. So that could be the HR department, could be the sales department, could be marketing. And that's the interesting thing about AI is it applies across all areas. So usually when I get involved with a customer or members of my team, we understand So, who are those executive sponsors that are really interested in helping transform their business? And usually we'll do a bit of education on grounding people on what AI is and isn't, and also address issues around responsible AI, because that always comes up. But then we'll shift to, well, tell us about your business. Of all the things that happen in your organization, what are the biggest levers where if you were able to increase productivity or quality of work, it would have an impact on your business because almost guaranteed. AI can be applied in any area and so start with the biggest problems that you have and then design AI into it instead of trying to understand AI and look at the hundreds of use cases that people have already published out there.

Heather:

So as we spin this over to just conversational AI specifically, with that being such a huge disruptor in the business right now, we've seen chatbots in the past, and they've been around, what, 10 or 15 years. So what's different about tools like ChatGPT, Copilot, Amazon Q? What's different about those? And what's taking those to the next level as compared to just a simple chatbot?

Craig:

Yes, it is different. And in fact, chatbots probably go back to 1966, where someone wrote something called Eliza, and it simulated a human conversation and was actually pretty good. And through the years, people have had different formations of, well, how can we simulate And the approach used was a rules based approach, meaning the programmer tried to anticipate all the interaction that might happen and would write code for all the different interactions. And then into the 70s and 80s, it got even more narrow and said, well, let's, can we simulate a doctor to do diagnosis? So they would talk to a doctor and try and encapsulate all the rules and these were called expert systems that tried to do diagnosis. But the challenge is you need to document all the rules and you might not get it right. And when things change, you have to go update the system. So the big change is, hey, instead of trying to document all the rules, Let's let the computer figure it out, which is exactly what generative AI and neural networks or deep learning is. It's let's get a whole bunch of data, give it to the computer and train it and let it figure out what the patterns are and how to behave. And so when you see conversational AI or a chatbot today, it's not a human that has programmed all the rigid rules behind it. It has learned based on being trained of all the data on the internet, how to have a conversation. And that makes it much more flexible, much more natural. It's fascinating when you combine it with synthetic voice. If, if you've tried a Google notebook, LLM and the podcast example, you'd realize, Oh boy, uh, is at some point these podcasts going to be all automated. And that really has changed the way that you can interact with computers. Because it has shifted from the rigid human programmed rules through to no, no, no, let's take a massive amount of data and let the computer figure out what it means to interact with a person.

Heather:

That's an amazing shift and it seems like quite a problem solver. I can't imagine. being diagnosed by a computer with a set of rules. As far as use cases, I'm going to do the same thing that customers do to you and say, Hey, what problems can you solve with this conversational AI? And maybe what has soft choice specifically solved? with some of these innovative applications.

Craig:

Sure. Often where people start is customer service or customer support. And so when you think of an organization, there's a set of customers that they serve. And right now there might be different ways to interact with them. Could be a legacy chatbot, it could be calling in on a phone, could be sending an email. And the opportunity to support customers in different ways is now available with conversational AI. So one example, we've worked with a couple of municipalities and they want to serve their constituents better. Taxpayers like you and I. And if you think about a municipality, people would have many questions and want to interact in different ways. Like, what hours are the library open? Or can I register my kid in soccer next fall? And the first step people think about is, okay, well, how do we take our existing support and just replace it with conversational AI? I think there's an opportunity to do things that Organizations have never been able to do before because of conversational AI. So if I take the example of a municipality serving its constituents, AI is very good at serving in different languages. So if previously you could only provide English based support for your constituents, all of a sudden you could expand that to many different languages. The interaction allows somebody to find information much faster. So sometimes people get lost in websites and documents to find the information that they need. I don't know if you've ever experienced that, Heather, but when you have conversational AI, you can just ask, Hey, this is what I'm looking for. What do you suggest? Or what is the answer? And what's particularly interesting then is if people are interacting with the information in this way through conversational AI, The AI can then also understand what are people asking. And so if you're the mayor, you'd be able to say, so what's on people's minds these days, what are they asking? And even sentiment analysis when they are asking what's the tone of voice. And so not only does it serve the constituents better. Because all of this information is being funneled through an AI, you can have an understanding of what's on people's mind and be able to react to it to address their needs better.

Heather:

I'd love some more use cases. How about insurance? Is there any applications there?

Craig:

Yeah, it's similar to the customer support example. Sometimes what we see organizations start with is that they're not quite ready to expose it to the end customer. Yet, they do want to service the customer better. And if you imagine that you're an insurance agent supporting someone, there's a massive amount of information that you have to know and understand if somebody has a question. And there might be very different manuals and, and different pieces of information. And so when you get an inquiry, you might have to search all these things to, to get an answer. So good, good way to bridge a full conversational AI system. directly with the customer, is instead to give the agent access to conversational AI. So, instead of them having to browse through all the information, they can say, this is the situation that I'm facing with this customer, and this is what their questions are, what would your recommendations be? And it would be able to parse through all the policies and information and surface an assessment of the situation. And then also provide some suggestions on how to continue next with the customer.

Heather:

As we look at both of those examples, I want to go back to that results driven impact that we have with conversational AI. So what were the measurements of success for both of those applications if you were to measure that and take a look at it and pass that information back to that end user?

Craig:

Yeah, so it's early days. So for organizations to actually be able to report back results, they're not quite at that point yet. But the metrics they are trying to address are the existing metrics that they would have today. And if you're in the insurance industry, for instance, customer satisfaction is key. So if you're able to provide a higher quality answer faster, likely the customer satisfaction is going to be higher. Well, guess what? That's exactly what the AI can do, is it can provide higher quality information faster. If you are a municipality, and one of your metrics is, how well are you serving your constituent base, and you measure that by number of inquiries, and if people are Inquiring about things, that means that they're interested in the services. If you're able to expand in different language and make it more accessible, you should drive the number of inquiries and therefore you're better able to serve your constituent base. This also works very well through an audio channel. So for someone who might be visually impaired or just not be able to access or navigate a chat bot, certainly doing it in an audio fashion where you just call a number and ask the questions, it works. Wonderful there too. And so it provides different ways to connect with the diverse range of people that you might be serving so that you can maximize a customer satisfaction. And then if you're a revenue driven organization, by having those different ways to access people in a diverse way, that means it could increase your customer base. And you can even personalize it even better. So if you understand that the person that you're conversing with is a 16 year old versus somebody who's 70 years old, the AI can change the way that it interacts with the person, including the language, therefore making it more personable and stickier with the end user.

Heather:

I'm just amazed by this technology and how it can be so useful. It really sounds like these organizations that are utilizing it have taken a look at Their key performance indicators, what makes them successful? And then you mentioned, it's not necessarily about tracking exact results. It's about asking the right question of what makes us more successful and how can we improve that? To get that business outcome that we're looking for. And so that's the tracking and the results driven aspect that really makes sense. So that's really cool. So if we were to look at any other industries, do you have maybe some novel ways that you're seeing conversational AI being used? I have heard through the grapevine that we have some great examples, uh, from Wendy's that we might be able to talk about.

Craig:

Yeah, Wendy's was one of the early movers. Something that almost all of us have done is go through a drive thru, and it's a perfect use case for conversational AI to interact with you ordering your food and asking questions about it. So they started with a pilot, just started with a single store, tested it out, expanded it to a broader set of I think 35 stores. and have an intent to expand it even more. And part of the reason that they did it was to give the servers more time. So not necessarily to remove labor from the restaurant itself, but to allow people to focus on what's important, which is preparing some good food to give to the customer rather than focus on the order taking. So it'll be interesting to see how this unfolds. Some of the other restaurants have played with it a bit, backed off a bit. I think there's still some caution on without a human in the loop what might go wrong, but I have a feeling as organizations get more comfortable and they see success and see their competitors be successful, you'll see a exponential growth rather than a linear growth of these types of use cases being deployed in the industry.

Heather:

I can't wait to see it. That will be great. So as we look at all of these things that organizations are seeing that helps them be more successful, as we're looking at the way that they judge their success rates, what are the biggest business benefits that you're seeing customers realize?

Craig:

Yeah, again, early in the industry. So people haven't been really able to report back. Here's exactly the dollar saved or the number of customers added. There was a study done by, I believe, the Harvard Business Review. that it gave a set of employees at the Boston Consulting Group generative AI and trained them on it, and the other set did not. And what they found was that group that had generative AI to assist them in their job were 25 percent more productive, but not only that, their quality of work increased by 40%. When do you ever see both productivity and quality go up at the same time? Usually they're opposite. Hey, we got to hurry and get this thing done and quality is terrible. Or, hey, we got to get this right, slow down. So there's been some early stats from a productivity and quality perspective. And we'll wait and see as organizations report back other types of metrics that change as they get these into production.

Heather:

Well, even with that Wendy's example, you mentioned That Wendy's was looking to give the servers more time, which I imagine would also make the servers happier. So there's your employee buy in, that happiness of your employee experience. But then again, if the servers have more time for customers, that's going to increase customer happiness and most likely productivity as well. So that's three areas of success that I could see coming out of a project like that.

Craig:

That's a good call out. And there is fear of. AI taking people's jobs. I think what is showing up more though is Employee experience goes up when AI is there because they're able to be more effective in their job and be able to deliver higher value results than what they were able to do before.

Heather:

So with all this success, I have to go back then to challenges and pain points because I've always got to ask those kinds of questions. So as a business leader yourself, what are some of the biggest challenges that you are seeing with developing and implementing this technology?

Craig:

Funny enough, it has less to do about the technology, but change. And so, getting AI into an organization is change. And so, you need to be able to have sponsorship, money, and time. And most organizations don't have an AI strategy. They are playing in a few areas, but they haven't written down, Hey, here's the project charter for what we're going to achieve. In 2025, here's how we're going to do it, the resources allocated, and the way we're going to track it at our executive level. That's the biggest challenge is organizations committing that this will be an executive led initiative, and we're going to put resources behind it, and we're going to have metrics to track it. What I'm seeing is experimentation right now. That's the same with any technology or change in their organization. If you don't have the resources and executive sponsorship, you won't achieve the results that you desire. I'd also say there is some fear of AI, how it might not give. The right results because it isn't rules based when in any computer program that's rules based you can predict the output, you know, very much exactly how it's going to work, but Heather, I'm sure you've played with chat GPT or a similar technology and you would put the same thing in twice and get a different results and it's because it's a trained model. It's difficult to understand how is it working inside. And that makes people uncomfortable. And so that can slow things down with organizations too. As I mentioned, often for customer support, organizations will start internally first so that there's a human in the loop and able to validate before they take the human out of the loop and deploy it directly to customers.

Heather:

That's an important step. I haven't heard that mentioned before. So especially when you mentioned the example of having a chatbot, that meant for customers of an organization but for the internal folks to be able to access data quicker. So starting out there seems like a great use case because you're getting to know the technology, you're getting to see how it works, you're increasing productivity, and you're not affecting, not necessarily directly, your customer base until you figure that out inside your organization. I've got to ask about security. I know we didn't really plan to cover security, so I'll keep it small. But like, as you're designing these systems and having these art of the possible conversations with customers, is security top of mind in these discussions?

Craig:

Absolutely. Generative AI, like if you think about Microsoft Copilot, it doesn't make you more unsecure. But if you have existing security problems, it will make them worse. That's because. Without it, you may not be able to discover that there's a security problem. As our CIO says, it's security through obscurity. But AI is very good at finding the information that you're permissioned on. So if you're permissioned on information that you shouldn't be, AI will find it and take advantage of it. So often a first step that organizations take is Okay, we need to do a bit of a security audit. Make sure that we got permissions locked down the right way on our documents and other types of information, which is something that organizations should have been doing for years. It just tends to be lower on the priority list, but AI has bumped it up.

Heather:

And as we think about all of these fears of the future and change and innovation, I'd love to ask you, and I know it's early days, what does the future application of conversational AI look like from your perspective in the near future?

Craig:

There's a somewhat predictable and then there's the, we don't know. And so it's obvious that any customer support function, this can do a really good job, if not better than a human. And so anywhere there's customer support, conversational AI will be involved in changing the way that those things are done. We will see the technology get better at an accelerated rate. So. If you look at the headlines, I did just some rough math. In the past month, there was over 50 billion announced of investment by Meta, AWS, OpenAI, and AI infrastructure. So what's happening today is like the laying of the railroads or the laying of the internet lines. The infrastructure is going in place. And right now it's the AI infrastructures being laid in place. The thing is, you don't know what's going to be built on top of it. So, we didn't know what economic and social impact did the railways have. You couldn't have predicted it, but it was huge. Where people lived, everything else, that really affected the world. Obviously, the internet completely changed how we communicate, created social opportunities, created social problems, opened up economic opportunities, and it was unpredictable. The same will happen with AI. We know areas where it will help, but it will be fascinating to see what people come up with on top of this AI infrastructure. To solve problems that we don't even know exist today and create new things rather than doing the same things better.

Heather:

Well, we don't want to do the same things better. We got to do some new things too. I could ask you a million more questions, Craig. I really appreciate your time. One last one. Where can listeners go to learn more about SoftChoice's conversational AI initiatives?

Craig:

Sure. Softchoice. com. You'll be able to see pretty close off the top page. We have some information on different aspects of AI and conversational AI is included in that.

Heather:

Wonderful. Thank you so much for your time. Thanks to leaders like Craig McQueen, companies across many industries are implementing conversational AI solutions that drive innovation, efficiency. and real business impact. This technology is reshaping business as we know it. As it continues to evolve, it's clear that the organizations embracing it today are the ones leading tomorrow. I'm Heather Hoskin, and this is The Catalyst. See you again in two weeks. The Catalyst is brought to you by SoftChoice, a leading North American technology solutions provider. It is written and produced by Angela Cope, Philippe Dimas, and Brayden Banks, in partnership with Pilgrim Content Marketing.