3277: Hexaware Reimagines Customer Experience Through Agentic AI

[00:00:04] What happens when AI is no longer just a tool, but an agent, a colleague that works alongside your staff? One that remembers, reasons, collaborates and even learns from every single customer interaction? Well, as businesses scramble to stand out in an experience-led economy, this isn't just a technical evolution.

[00:00:27] It's a cultural and operational shift, one that touches everyone from the contact centre to the boardroom. So I've asked my good friend Bennett to join me from Hexaware, and together we're going to unpack the emergence of agentic AI and the profound ripple effect that it's beginning to have across customer experiences, employee roles, and also how businesses structure their operations. So Bennett is going to be sharing his new breed of AI,

[00:00:57] how it is making it possible for agents to serve customers in any language from any location and even identify sentiment in real time. And in doing so, steer conversations with greater empathy and indeed precision. I will also examine the psychological shift needed from customers and staff alike to embrace this next wave.

[00:01:21] Yeah, the tech is cool, but change management becomes just as important as the tech itself. And as Bennett will share with us today, preparing for agent-to-agent interactions between your personal AI assistant and the enterprises is no longer things of science fiction. The future is now. So, what does all this mean when AI can act on your behalf with autonomy? Is your business ready for a world where the customer doesn't just call,

[00:01:49] but sends a digital assistant to negotiate, request, and resolve in real time? And how do we design trust into this future right from day one? Well, it's time to find out by officially introducing you to today's guest. So, thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do? Yeah. Hi, Neil. Thank you for having me on the podcast.

[00:02:15] My name is Bennett Kumar, and I'm part of the leadership team at Hexaware Technologies. Hexaware, as you're aware, is a large global transformation-focused AI-first firm servicing customers in multiple industry sectors globally. And I run a business unit within Hexaware, which focuses on operations, transformation, and services.

[00:02:38] So, my team works with the business functions of our customers to help them to get more value from their business operations. And a lot of the AI-based transmission interventions are very pertinent to what my team does for our customers. And I've been in Hexaware for 15 years now, and I've enjoyed every moment of it. And it is an incredibly exciting space that you're working in there.

[00:03:05] Gartner last year famously said that 2025 will be all about AI and AI agents. And quickly, the months that followed, there's been a lot of hype this year around AI agents. So, just to ensure we don't get anybody left behind here and put it in a language that everyone can understand, how do you define AI and why do you think it's such a pivotal shift for the future of customer service?

[00:03:31] Because I think a lot of people hear the buzzwords but don't fully understand what it means or what it could mean to them. Right. So, before Agentic AI, there was automation in operations. And we used to call it robotic process automation several years ago. That evolved. But the key was that a lot of that was kind of hard-coded, so to speak. So, you had a specific function and a specific activity that needed to be done.

[00:04:01] You had to configure that to the automation platform, whatever platform you were using. And it'll go and keep executing that same task over and over. But we are all aware of the advent of AI, Gen AI and all of that. So, given the wide adoption of AI and increase in capabilities of AI, that is now getting into this core automation engine that people have built.

[00:04:29] When those two get combined, you get Agent AI. So, basically you get an AI agent which can keep in-memory business context in which it has operated in the past. Can use that context to plan and execute a set of tasks autonomously. Can execute those tasks and learn from it. Can you do that and then keep improving? And can seamlessly collaborate with other AI agents who are tasked with doing similar or varied tasks.

[00:04:59] And they all can come together to deliver a business outcome. All with eventually minimal human intervention and supervision. So, we are all familiar with the concept of agency, right? Agency is the ability to do something, correct? You are empowered to do something. So, this is where a piece of AI is empowered to do a specific task. And can learn from it and execute. And so on. And this has tremendous implications for how business operations will be run.

[00:05:30] Because business operations today are run largely by people who are doing a varied set of tasks across the value chain. Some of those tasks have gotten automated so far through traditional forms of automation. But at the advent of agents, significantly higher number of tasks up the value chain will also get automated through these agents, AI agents if you will. So, yes, I hope that made sense. Yeah, it really does.

[00:05:58] For business leaders that are doing things the right way and thinking about business outcomes, ROI, value, measurable difference. When they're thinking about all those things before the technology rather than after. What would you say are some of the most immediate changes that a Gen TKI could bring to the customer experience? Particularly in terms of expectations and personalization. Right.

[00:06:22] So, the ability to dynamically alter experience, transactions and workflows is at the core of personalization. So, Neil, if you call somebody, you would want them to talk to you in a particular way. If I call that same number, I would expect to be treated differently based on what my preferences are.

[00:06:45] It's very difficult to do all of that dynamic routing and have a lot of that intelligence and have enough of a context of who Neil is and who Bennett is before the agentic AI revolution. But now you can really create AI agents that have memory, that can reason, that can learn from all the past interactions with Neil or Bennett and can really drive a hyper-personalized experience. Right.

[00:07:13] And I'm not even going into other aspects of translation and human-like communication and all of that, which is all the interface channel. But how you communicate changes quite dramatically and can be hyper-personalized. And a lot of that personalization depends on data that comes from underlying enterprise systems. And in the AI world, it is now possible to establish very robust data pipelines to underlying enterprise systems.

[00:07:44] And agentic AI has the ability to make API calls to work off of underlying enterprise data systems to really drive that hyper-personalized experience, if you will. Yeah. So I see that as being a big difference in terms of how this will evolve in the future. And before you came on the podcast, I was doing a little research and I was reading how you previously mentioned that live voice translation powered by Gen AI could be coming close.

[00:08:11] How close do you think we are to that becoming a standard in customer support? And what does it mean for global service teams when that moment does arrive? We are very close to that moment. In fact, if you ask some folks, they would say the moment has already arrived because there are several use cases already in practice. The key is to avoid latency and to make it near real time as possible.

[00:08:37] And to be able to drive translation across different channels, right? Somebody calls voice channel, it gets converted to a text and can be handled as a chat interaction by a human agent sitting anywhere in the world. They can then respond in a chat interface and then the AI voice can respond back in the right language. So all of that is already happening. Is it happening at scale? Probably not yet, but we're very close.

[00:09:08] We're very close. And what this means for global services teams is that you really can have the kind of anywhere delivery model. I mean, onshore, offshore, those concepts can quickly disappear because you can now have the teams with the best expertise on the business process.

[00:09:27] So, you can see that there are other AI-enabled software that also neutralize accent. So even if a human agent has to be in the loop to do this work and they speak the language that the customer is speaking in, their voice and accent can also be neutralized. So which means that we really can operate customer experience from any part of the world.

[00:09:57] And I'm talking about the remnant customer experience, assuming agentic AI and other forms of digitization take over a significant portion of remaining human labor. Can technically be placed in the location best suited from an expertise standpoint. You don't have to place those humans in a location just because they can speak a language. You can place them in a location where they really know the process. Then language becomes secondary.

[00:10:25] And we are already seeing that happen in some of our customers. It also means that you can move operations to locations that are not as expensive. From where they have to operate from a tier one location, you can operate from a tier two location because voice and accent can be neutralized. It has a lot of location strategy implications for service providers.

[00:10:48] And of course, when implementing any new technology or new tech project, many businesses often forget or neglect the importance of people and culture within an organization that are very often required to bring that technology to life. So what kind of change management do you think is required to help customers and indeed employees become comfortable with interacting with AI agents, especially in sensitive or high stakes scenarios? Right.

[00:11:18] Right. So it's my view that as with any form of change management, communication is everything. Answering the what's in it for me question is super critical. You have to get both of that right. From a communication standpoint, humans should be told where appropriate that they are interacting with an AI agent. It has to be directly told. There could be a lot of AI agents in the back end doing a lot of back office tasks. The humans need to be aware of that.

[00:11:47] But at the interface layer, if there is AI involved, humans should be made aware that they're interfacing with AI. Therefore, there is no gap in what they expect. They shouldn't be expecting to talk to a human and talk to an AI instead. And that causes an issue.

[00:12:01] But at the same time, the communication should also include the fact that there are guardrails in place, which ensures that there is a level of protection for the human in terms of what the AI can and cannot do. So those guardrails have to be possibly mentioned. It cannot be mentioned in every conversation. It has to be mentioned as part of broader customer communication.

[00:12:28] When you roll out an AI agent, the customers are going to be interfacing with the communicator that it is safe to use. It has all the right guardrails that things possibly will not go wrong. They should feel comfortable. Then they should also quickly get to the point of what's in it for them. Right.

[00:12:49] And it has to be covered in the communication that through AI, you're going to get a hyper-personalized, faster, more accurate, anytime on service. So service quality and your satisfaction levels are going to go up. Provided you adopt this channel. Right. So you have to give them a view that what's in it for them. But at the end of all this, there has to be a fail safe.

[00:13:15] And they have to have the comfort that if they're not comfortable with interacting with AI, that there is a human available for them to interact with. For certain types of interactions, they can always escalate it to a human. Right. So you should be able to do that.

[00:13:34] So basically, great communication and answering the what's in it for me question for customers is important while reassuring them that there are guardrails in place to ensure that their personalized data is protected, their interactions will not be misused and all of that stuff. The reason why we all distrust AI is because we think it'll go out of control. We should reassure the customers that things are in control. Yeah. Completely agree.

[00:14:04] So important for the customers to understand what's in it for me. And I think also the staff in an organization, they need to be taught as well. What is in it for them too? Because many will automatically think that, hey, this is coming to replace me. That's the kind of narrative that they may have seen on their news feeds where, of course, the real magic happens when the two are working side by side. So on the operational side of things, how does Agentic AI empower those customer service staff rather than replace them?

[00:14:33] And also what new roles or responsibilities are you seeing emerging as a result? Right. Right. So I will answer this in two parts. So what is the enablement part and then the roles evolution part? We've all had this North Star and customer experience that you need to provide a single pane of glass to an agent across all the systems that they work in. They should have access to all of it seamlessly. Right.

[00:15:01] In many legacy environments, including regulated industries like banking and insurance and health care, a customer service agent typically needs to toggle through five or six different systems if they have to answer a question. So folks are focused a lot on trying to give that single pane of glass view to an agent so that they have all the information on the fingertips. And a lot of that was kind of hardcoded or connectors built enterprise systems and all of that. And it's in place. What you now get to the agent is that pane of glass can evolve.

[00:15:31] It needn't be hardcoded. So if your business logic and your rules change, you want to add new types of customer experience, you want to add even a new type of service, answer 50 new questions. Right. You don't have to do much change in the underlying integration and the plumbing to continue to provide that highly contextual and relevant data to the agent. Right.

[00:15:58] So it really enables agents to have the right level of information. At the fingertips all the time. In a very intelligent way, it can also help them to really. Drive customer transactions. In a highly empathetic manner.

[00:16:21] The AI can provide intelligence in terms of real time sentiment analysis and the like to give them a view of what the customer is feeling. And also related back to what the customer has reacted to in the past in similar situations so that the agent can drive a very effective conversation.

[00:16:43] A lot of the times the frustration of an agent and why people burn out is because they just don't like talking to customers who are rude to them. They don't know what to do. And AI can help that in a big way. AI also can help more intelligently route and manage workloads. All of that. AI enabled workforce management and all of that.

[00:17:05] And therefore, you really can provide a way better system of allocating work to agents so that you can sense burnout. You can pull work out from certain agents as needed. You can provide them more opportunities to recover from a tough conversation. So all of those is AI led operations management that can make the agent's life much, much better.

[00:17:29] And of course, the classic use cases of agent assist is that you have a chat or a voice interface that an agent can rely on to quickly check on certain things. Place the customer on hold and then you ask your AI assistant. Almost like a personal AI assistant. Hey, this customer is asking me this. What do I do? Where do I get this information? So all of that is in play. We're also seeing roles evolve and change.

[00:17:55] And the biggest change we see is the management layer now has to evolve to managing humans and agents working together as opposed to just managing people. And that requires a level of technology savvy. It requires an elevation of your thinking because sometimes human nature is, hey, if I manage 100 people and you manage 200 people, you are the bigger guy. I'm lesser than you because you manage more people.

[00:18:24] In the AI world, it's not about number of people that are important to me. It's really about the volume of work that I'm able to deliver or business outcomes I'm able to deliver. That requires leadership teams to change in their thinking. And therefore, we see an evolution in terms of operations leadership.

[00:18:45] We also see changes in the agents' roles themselves in terms of being more upskilled in their capabilities as opposed to just doing routine work. And not everybody can upskill. And therein lies the thing where people need to learn and adapt to this new work environment to work alongside automation. So we see that changing. We see AI impacting how you hire and train resources. We see a lot of AI application in the training function.

[00:19:15] And then how we hire, how we drive human resource capital. How do you drive workforce management? So all of those roles are getting impacted by AI and becoming AI-enabled. And therefore, you need people who are capable of working with AI. So kind of long-minded, but hopefully that gives you an answer. Yeah, it really does. Well, thank you.

[00:19:36] And I'm curious, when you look out there, what industries do you think are best positioned to adopt agentic AI and do it at scale? Where do you expect the most disruption? So I would say any of the consumer-facing industries where there's not a whole lot of regulatory overhang will be the first to take advantage of this opportunity.

[00:19:59] Retail, e-commerce, travel, transportation, hospitality, where there's not a whole lot of regulation, but they are highly business-to-consumer industries is where you can see a lot of adoption. And then I would think that the regulated industries will follow subsequently, banks, healthcare, insurance. There are certain processes within these regulated industries that will lend themselves to agentic AI, right?

[00:20:28] Like the first notice of loss process in an insurer or just submitting a claim in a healthcare provider or scheduling an appointment with a healthcare provider. Those are all things that can potentially benefit from agentic AI. Similarly, in a banking scenario, reporting a lost card, things like that. A lot of routine transactions that can significantly be transformed by agent, but there'll also be a lot of processes where humans will still need to be in the loop.

[00:20:56] Like if you're processing a payment, a lot of it can be automated, but above a certain limit for certain types of transactions, you would need human validation, right? Things like that. Things like that. So the business-to-consumer industries first, followed by regulated industries next in some measure is how I see it. And what about yourselves, Hexaware?

[00:21:18] How are you approaching the integration of an opportunity of multilingual AI-powered tools that can support a more diverse customer base across regions? Because it really does feel like an exciting opportunity here, one that we don't talk about enough. Right. So we at Hexaware are AI first in everything we do, and that has been our stated position for a bit now. And that continues even in this space as well.

[00:21:44] Our stated position is that every new opportunity that we drive in the customer experience world will be agentic AI first. So we look to first understand what the current state is. We look to architect an agentic AI-based view of the future, like a target state model. And then our entire organization is geared towards taking customers towards that target state.

[00:22:09] We seldom start with just, okay, I'll just run some customer experience staff for you, and then I'll drive transformation later. So that model is over for us. I mean, we are looking at, I'll work to define an agent-based future state for a customer and then help them to get to that future state. And then we are also putting skin in the game, committing to agent-based outcomes so that customers can hold us accountable to get to those, get to that end state.

[00:22:37] So to that extent, we have a portfolio of customer experience agents designed for specific industries. We have agentic in various domain processes, back office transaction processing in banks, insurance and healthcare and other sectors. We also have built a lot of technology agents, be it in data, in development and many other areas where agentic AI can play a big role in driving transformation.

[00:23:05] So it's a holistic approach, cutting across technology and business operations with a commitment to getting a customer to an agent-based future state. So that's how we have kind of tightly integrated this into our value proposition. And to bring to life everything that we're talking about here and help business leaders understand exactly where we're heading much quicker than they might realize as well.

[00:23:32] What's your overall vision for the role of personalized AI assistants in customer service? And how should businesses be preparing for that next evolution right now? Because it's already here, isn't it? We're not talking about the future. The future is now. The future is now. When you say personalized AI assistants, I would think you're referring to AI assistants that be available to you and me as consumers, you interact with businesses. Yeah, that is really the big trend that's going to drive a lot of this.

[00:24:01] So like we could be having this conversation, but I could have an AI assistant processing my insurance claim. I could have just told it to do it for me on the side. And it's talking to Geico. It's lodging a claim. It is negotiating what's going on. They might be asking it some additional documentation, photographs, what happened. It could be summarizing all of it and sending it to Geico as an example. I'm just using that as an example.

[00:24:26] When we have personalized AI assistants, the volume of transactions that organizations need to be prepared for will significantly go up. Because if I had to call a customer service line, I'll probably call once or twice and give up. But my assistant is not going to give up. My assistant is going to interact, try again, maybe talk to an AI agent at the company's end who will then respond.

[00:24:52] So there'll be an agent to agent interaction between my personal AI assistant and the enterprise's agentic AI. That'll then kick off a slew of other autonomous agents that'll try and service this request. So it is going to make life a lot easier for consumers, personalized AI assistants for a whole range of things. Not just to do research or to play the latest song, but execute real transactions.

[00:25:21] It's like having your own office assistant in a virtual world. And that's going to significantly bump up transaction volume for enterprises. And they have to be ready for it. Because if they can't service AI initiated transactions, then they'll be at risk of losing market share. Well, thank you so much for sharing your insights with me and everybody listening today. And before I let you go, I'm going to ask you to leave one final gift to the listeners.

[00:25:50] I always ask my guests to either leave a book that means something to them or they would recommend for our Amazon wishlist or a song for our Spotify playlist. All I'm going to ask is what would you like to leave everyone with and why? Right, right. I mean, I'll go with the song and this could kind of date me a bit, but that's that's fine. My favorite band is Dire Straits and my favorite song is the Sultans of Swing.

[00:26:16] And the reason I like the song is it really is teamwork in action. It talks about a band that just goes about doing their job. You have a guitar player, you have somebody who just plays rhythm, the people who play the different roles and they all come together and they create magic through music. And that's always stuck with me as a great example of a team that works together to create magic. And it has kind of stuck with me since childhood.

[00:26:46] It's like my favorite song by far. And then the man, he steps right up to the microphone. And I had to stop myself from playing air guitar, tying something around my head there. But an absolute killer tune with pleasure. I will add that to a Spotify playlist. But for anyone listening that wants to find out more information about Hexaware and everything we've talked about today, where would you like to point everyone there? You can just go to our website, hexaware.com.

[00:27:12] And I think that will then direct you to all the resources that are available, including all the blogs you've written, all our articles and our point of view on AI. And all of that is available on our website. I think that's the best resource. Yeah. There was so much I love from our conversation today. From, yes, the technology, what it means, what it will unlock. But also, what is in it for the customers? What does it mean for the customer service staff?

[00:27:38] How is this AI-powered voice assistant going to improve customer satisfaction? So many real-world examples. And you even left us with an absolute killer tune. So thank you for joining me today. Really appreciate you, Tom. Thank you so much, Neil. It's my absolute pleasure to be on this podcast. Thank you. I think this conversation shows. Agentic AI is not about replacing people. It's about redefining how they work and what they focus on.

[00:28:03] From empowering staff with intelligent assistants to removing language barriers entirely. The potential is enormous. But only, and repeat only, if we bring both customers and employees along for the journey. And I think Bennett's call to rethink how we train, how we lead, and even hire in this AI-first world,

[00:28:26] is a timely reminder that the tech may be advancing fast, but the human element is still central to its success. And with personal AI assistants already beginning to handle tasks we used to consider too complex, cumbersome, or sensitive, the question now becomes one of readiness, not possibility. So are we designing our AI experiences for scale and trust?

[00:28:51] And are we ready for a time when your customer's first point of contact might not be a person at all? Let me know your thoughts. How do you see Agentic AI reshaping your world? What do you like? What do you not like? What excites you? What concerns you? TechBlogWriterOutlook.com LinkedIn X Instagram Neil C. Hughes So much to think about. So I'm just going to sit in my favourite chair, pour myself a whiskey,

[00:29:18] maybe create a makeshift headband and play air guitar to sort of swing. I think that's what Bennett would want. And I want you to hold that image as I walk off into the sunset. I'll be back again tomorrow with another guest, but thank you for listening as always. And I'll speak with you soon. Bye for now.