AI For Improved Customer Experiences – Techstrong AI Podcast EP 30

Hello, and welcome to the Techstrong AI Podcast. I'm Amanda Ani, and with me today is Chris Ro. He is the Vice President of Technology and Innovations at iTech Ag.

How are you? Good. Thank you for having me.

Amanda. Can you share a little bit about the company, first of all, and your background before we start? Yeah, sure.

So as, as you mentioned, my role in iTech is the Vice President of Technology and Innovation. Um, iTech itself is a, a technology consulting firm. We help our clients identify, enhance, and streamline business processes, uh, through integrated technologies in the solutions.

Um, in my current role, I have the responsibility of overseeing our portfolio of technology focused, uh, solutions and projects across our customer base. And, you know, I think relevant to today's topic, um, here at iTech, right, we're increasingly integrating AI capabilities, uh, within our solution platform so that we can intelligently automate workflows, uh, and interactions across, uh, the enterprises that we support. Um, I guess by way of background, you know, I've always been a, uh, technology enthusiast, uh, highly inquisitive about innovation and, and emerging technologies.

I've worked in the technology, uh, consulting space now for about 20 years. I started in cybersecurity, involved my expertise into, uh, wider IT operation capabilities, uh, in support of enterprise business applications through, you know, various different business lines such as financial management, human resources, customer relations, uh, for security. And, uh, excited to be here today.

Uh, appreciate you hosting me on the podcast. Looking forward to our discussion. Wonderful.

Well, it seems like when it comes to integrating technologies, almost every space is looking at how to harness AI technology, which you mentioned. So what are, maybe what are some good use cases where you've seen companies have a, um, a good return on investment? And then maybe we could talk about, um, the process of implementing ai?

Sure, sure. Yeah. So I mean, within the federal space, right, not just, uh, the customer experience, AI is, is increasingly adopted, right?

Uh, we're seeing it for all sorts of purposes from chat bots to using generative generative ai, uh, to generate code, um, right starting basis for code, um, document processing. There's tons and tons of forms in the federal government. Uh, so translating those handwritten forms into digital and then, you know, uh, uh, pasting those over to, to various fields within, uh, the system repository.

Um, data analytics and trend analysis is huge used in security. It's used in fraud detection. Uh, it's used in just making business decisions around trend analysis.

Uh, I mean, we're even seeing it used for, for traffic management, right? Uh, with peaks and in, in valleys and, and different, uh, traffic patterns given the time of day and, and rush hours. So it's, it's really being used everywhere.

Um, where we intersect with it more often than not is, um, and this is, uh, in relation to, to the customer experience, um, is, is chat bots, right? When, uh, you're interacting with, with customers, now we're seeing a really high return on investment, um, by putting that chat bot kind of as the tier zero, uh, agent, letting that, you know, self service, uh, the requester, hopefully they can get the, the, the data they're looking for, they can get the answer they're looking for through that chat bot, if not, and escalate to the next tier. Um, and even using AI at the next tier, when, when I may be interacting directly with the human, we're seeing, um, a lot of enhancement in the personalization of that experience, right?

So much data is gathered, uh, and continue to gather from different sources and even through, uh, historical interactions of that customer. Um, that, that really gives us a lot of data that's usually presented on the screen, uh, even transcripts or scripts, right? Uh, suggestions on how you interact with that customer, uh, or how you answer their questions may be presented in front of you.

So it's really, really making, um, these customer interactions very, very efficient. And, you know, it's allowing us to, uh, often not reduce the, the footprint, uh, of the resources, providing that capability, but rather empower them, uh, or enable them to tackle bigger, um, and more, more critical or, or difficult challenges that AI can't handle and, and the human touch is needed. So, um, that, that's really where we're seeing it most.

I, I know there are tons of other areas, um, that, that, you know, realizing, uh, significant cost savings, efficiencies and benefits, but, um, again, where we intersect with it, where we're, where we're using it in our systems and our solutions, that's probably, uh, the biggest area we're seeing. And so, not only is it helping right then in real time, but I imagine companies are able to, with all that additional data, they're able to make changes across the board, um, to their websites and to the way they run their, their customer experiences. Yeah.

No, that's exactly right. Right? Like getting back, I guess to the, um, that the data analytic that have been trend analysis piece, right?

So leveraging trends from customers, right? Yeah. I mean, you can see where a mouse is at any given time on the screen.

Um, so kind of knowing maybe what pages they're interacting with more, or where their eyes are being drawn to, um, where we're seeing maybe, uh, routine data entries being incorrect or, you know, not the right products. So we can kind of gather, um, those inputs, uh, where there may be pain points where we can streamline processes from that data that's being collected and, and leverage ai, right? To kind of help us figure out how do we, uh, how do we make this more human-centric, right?

Human-centered design's a, a, a big, uh, concept in the federal space right now. So how do we maybe make this more, uh, easily interactable and more aligned with, with human-centered design principles and just ease the interaction, make it easier on, on the customer, and more intuitive. So for companies that are looking to implement AI technology for chatbots or any sort of, um, customer experience enhancement, what is step one, and then what are some of the roadblocks that you see business leaders coming up against tips for that and how to handle those?

Yeah, sure. So you, you know, whether it's a chat bot or, or another technology, right? Um, we've recently helped, uh, implement a solution where, uh, one of our customers was going through, um, uh, a translation or, or conversion, right?

From physical documentation to digital documentation and using ai, um, you know, they had all those, those handwritten form scan using ai we're able to generate all of the, the digital, uh, counterparts to them and, and, and build those systems out. So it doesn't have to be chatbots, but, um, you know, I think along those lines, the very first, uh, you know, step in, in getting to your first question would be where are you gonna get the biggest bang for your bot, right? Where's your, your, your business gonna benefit most from the implementation of ai?

Um, and how am I gonna get that, that return on investment? Um, right? So like anything, you know, it's great if I can, can implement ai, but if I'm only getting a 50% return on that investment, uh, and I'm losing the other 50%, right?

Then, then that's not a sound business decision. So, one, making sure it's right for your business and the, uh, business challenges that you're gonna tackle, um, right, where, again, where are we gonna get the big bang for the buck, and what's the right solution to do that? Um, you know, once you're there, it it, like anything, we're gonna evaluate, uh, the different, um, products that are out there to fulfill that solution, right?

So, you know, there are a lot of different chat bots out there to select from. Um, they all have different pros and cons in terms of different knowledge bases they can hook into, uh, knowledge bases. They may be pre-configured with, um, you know, learning models that they've been pre-configured on.

Maybe they're, so that's more, uh, customer service fit based for, um, like, right, uh, e-commerce versus one with customer service phase for IT service management. So, right, figuring out what, what the right product is, um, and then when, when implementing it, right? Like being able to pull through that chat bot to, uh, the page, right?

Because it's often overlaying on a page that, how am I integrating the chatbot technology with my, um, you know, service management technology. And then really the biggest thing is what are, are the learning models that it's, it's attached to or, or trained on, right? And then what are the, um, knowledge bases it has access to it, it can only answer and help, uh, from the chat bot standpoint, right?

Is as well as it has something to understand. So, um, one thing we're seeing a lot of right is, um, a lot of customers, right, and especially in customer service, have their knowledge bases, uh, whether it's SharePoint, whether it's knowledge articles in ServiceNow, whether it's something in Salesforce, right? Whatever it may be.

And, um, you know, connecting it to those knowledge articles, but also allowing you to access, uh, like the user profiles and user data within that system. And that's where we get to start seeing, um, the trend analysis, right? Like, what, what is this user's history of interaction been with us?

What tickets have they previously opened? Are there common themes? And, you know, this user keeps having the same problem and the same answer.

Here's how I open answer it. So it's getting the right data. So I think, right, the biggest things are, um, understanding where you're gonna impact your bus or make the greatest impact on your business, selecting the correct technology, uh, building the proper integrations between your existing technology and the, the, uh, the AI solution.

And then arming that AI solution with learning models and, and, and data to reach back into. Um, so that, that, that would kind of be my recommendation when looking to, uh, implement ai. And again, this is just specific to, to the chat bot model, but I think it's, it's really applicable to, uh, you know, any piece of AI that, that you may be implementing.

And of course, from some things I've read and heard, I would say that business leaders do need to be careful though, because I know I saw something a few months ago where an airline was using an AI chat bot and it answered a customer wrong, and this turned into a big lawsuit because it cost the customer money. Uh, and so then they said it was, they weren't to blame, it was all the AI chat bot's fault, but that's just a technology. So what are your thoughts on this as far as, you can't just let the ai, you know, do it, you'll just expect that it's going to do everything perfectly and not take any blame for it.

Yeah, yeah. I, I'm not familiar here in that specific instance, but, um, certainly it is a concern. Um, I think the way we're seeing industry handled more and more is, um, there's something called human in the loop, right?

So it's AI not operating autonomously, but some level of oversight or even participation in any given workflow of a user, um, that may reduce the speed of the AI and the speed of getting answers to a customer. Uh, but I think great, when the newer the technology is you, I don't wanna give it the quote, quote, training wheels, and that's where the human in the loop comes. Um, and over time, it, it does get better and it does perform more So, uh, in this particular instance, right?

Maybe it's just a matter of taking the training wheels off too early, but certainly there are are big concerns with implementing ai, um, ethical concerns, right? What's, um, are there biases in the AI and the, the learning model and the data that that's being fed? Um, there security concerns, you know, AI has, has access to all of this data.

What are the, the screens or filters preventing it from, Hey, you know, I'm interacting with it, but I, it thinks I'm Amanda and it's giving me all of Amanda's information inadvertently, you know, leaking that data for kind of a, a new vector. Um, so yeah, ethics, privacy, data privacy, uh, those are the bigger ones we're seeing right now. And actually, uh, in the federal space, they, they, they are facing it head on, I believe it was back in March, that's, uh, the White House or RB issued a memorandum, um, for heads of executive agencies on how are they going to, uh, handle ai, right?

Really strengthening governance around ai. And, uh, you know, I I I think, not just that the federal government, but you know, industry altogether is, is really taking that to heart and it's really putting a framework around it, right? Designated AI officers who's gonna own it, just like we have now chief information security officers or, uh, chief Privacy officers, there are now chief AI officers, um, you know, building governance bodies within the agency.

Obviously it's just not one person making the decision. It's a collective, uh, org or group within the organization, compliance plants. Uh, and something I think is really cool, I don't know if you're aware of it, Amanda, um, but it also requires all the agencies to inventory that use cases for how they intend it to use AI and post that publicly.

gov, there's actually a, a, a whole website out there with a wealth of information around how the government's using ai, how they're handling some of these, these concerns and challenges with ai. You can actually see all the use cases posted out there, agency by agency, by agency. Uh, and I found that really interesting, right?

Um, just seeing, you know, the, the vast adoption and how quickly that's happened on or for AI and really just kind of taking the time to look through some of the use cases. There's some really cool stuff out there. So, um, you know, for your listeners or listeners if, if interested, I would encourage 'em to go take a look at that.

Um, 'cause it really does, you know, provide a lot of transparency to, to how AI is being used in the government. Yeah, absolutely. I'll have to go look for that.

And plus I like that that sort of provides a bit of accountability as well. Yeah, Absolutely. So what is one key takeaway you can take and give our audience today in regard to AI and the customer experience?

Yeah, right. So I think the, the biggest thing is, um, AI is, uh, amazing, um, advent that, that we're now seeing some of the benefits of it are astronomical. We're just seeing huge investment or returns on investment.

Um, but specific to ai, right? I think or, uh, the, the, the human experience, I think we always have to take into account, um, the, the human element of it, right? AI is, is not human.

It is gonna slip up. It, you know, you can, you can often tell when you're talking to AI and not a human. Um, so I, right.

I, I think the emotional intelligence aspect and empathy is, is something that AI can't provide. So the human touch is always gonna be essential when, when kind of managing the, the customer experience. Um, so, you know, I think the biggest thing is lean into AI as as much as you can because you really are gonna see, um, that the benefits of it.

But you have to caveat that and wrap all that with the understanding that at some point I need a release out of that AI interaction to a human in the loop. Uh, and, and, and, you know, insert that human, um, emotion, empathy, intelligence, right? When, when appropriate, um, 'cause you, you can never leave the human experience to, to ai, it's just not gonna be there.

But, uh, it is making really, really impressive strides. But I think you always have to have, um, an understanding and awareness. It's again, always dealing with, with, with the human touch that that needs to be baked into, uh, to the customer experience across board.

Yeah, that collaboration is important. It will be interesting to see as fast as this technology is advancing what's in store a year from now, what are we gonna see? It Is.

Yeah. Yeah. I think, um, you know, to the point I just made, I think right now we're interacting, uh, still interacting with, with, with phone calls, right?

And it's push a button for this, that, or the other thing. And, and it's almost counterintuitive to the point I just made, but I think we're gonna start getting to the point, right, where you're, you're talking to AI and Right, we have Alexa and things like that now, and you're actually going through a whole customer service experience, whether, you know, be likely via the phone. Um, and it'll be fully conversational, it'll be fully ai.

Um, but again, they gotta figure out how to have that release and, and if a human needs to step in how to do that elegantly. Yeah, absolutely. Well, I wanna thank you for coming in on our show and sharing your insights with our audience today.

Yeah, absolutely. Thank you for having me, Amanda. I, uh, really enjoyed our time and I look forward to, uh, hopefully talking with you again soon.

Yes. And thank you to our audience. Stay tuned.

There is more.