Terry Gerton You have worked inside the IRS and now outside the IRS, but always with a focus on customer experience and customer service. In that timeframe and from those perspectives, what do you think the most important lesson is that agencies have learned about customer support?
Kevin McCreight That’s tough because there are so many different places you can go with it. Terry, you know, I think customer service or customer support has really evolved into meeting the customer where they want to be, the channel they want to be at and first time resolution or close to first time resolution for them. And, you know, with a lot of agencies, they’ve had legacy channels, phone, walk-in. And the agencies want to go to the new modern channels to web, to mobile and phone self-service, but the customers don’t always want to, so there’s a give and take, so it’s having to offer that continuum of service, the channel that the taxpayer, the citizen wants to use and giving them that resolution. So if you’re forcing them to use a channel, make sure, or encouraging them to use a challenge, I should say, make sure that they can get the answer they want and not get frustrated by then having to switch to another channel. So if they’re on the web and they don’t have the service, do I have to call now? Do I have to go to a walk-in center?
Terry Gerton That is such a great baseline, especially now let’s come back and talk about customer service at the IRS because there’s been a lot of concern there. They’ve had tremendous workforce reductions over the past year. What are you looking at, at the IRS to tell you whether or not their customer service is actually improving?
Kevin McCreight IRS is always measured by, the biggest one is on the phone systems, is how soon can I get an answer? Is my call to answer? How long am I waiting on, waiting to get my call responded to? I think with the IRS and where they’re going, is to increase the self-service capabilities on the phone channel, to be able to provide more transactional self-service, not just FAQs, but real-time conversational and be able make account updates. And they, we have since 2019, when I was inside working to make those services more readily available on the phone channel. And they are continuing that effort and accelerating it to a certain extent to be able to use the modern tools, AI and others to increase that self-service capability.
Terry Gerton If those are the outcomes that agencies really care about, then where can AI make a measurable difference and maybe where does it not help so much?
Kevin McCreight It will make a measurable difference when the IRS is able to fully authenticate taxpayers in the voice channel. Meaning there are several different authentication methods. On the digital channel, they use ID.me and that allows the taxpayer access to self-service their full account. I don’t have something similar to that yet on the phone. So being able to do that can unlock a lot of other potential services and unlock faster service with agents. The other would be opening up more of the backend systems and being able to make self-service account adjustments. You know, the IRS has a lot of FAQs online. They have it through the phone system. But when taxpayers are calling, They really have an issue with their account. They want to be able to update something. They want them to be able to get a specific piece of information. And being able to provide account updates in the self-service channel, voice channel is really gonna move the needle for the IRS. I do wanna add, and I’ve been thinking about this, we’ve been focusing voice on the website. IRS has made huge strides in self- service and what they provide the taxpayer. And that’s a huge channel for them and really provides a great service. The voice needs to catch up.
Terry Gerton Kevin McCreight is vice president for account management at TTEC Digital. So Kevin, when you think about the improvements that you were just describing, what is the hardest part about integrating AI into a large citizen service environment like that at the IRS?
Kevin McCreight It’s resistance to change. It’s resistance to change largely on the people that are providing the service not necessarily a technologist but those that are responsible for providing that service to the taxpayer. When they’re doing it through a human channel there’s training that they have hands-on with. There is QA of the agents providing that service so there’s a great deal of certainty that the right answer is being provided, and it’s being refined. With the new AI channels and AI capabilities. You know, there is a huge spectrum of, you know, what it can provide, can provide the same types of service, but there’s reticence because do I trust that it will provide the right answer? How do I know it’s not going to give, take the taxpayer down a different path? We talked about training and QA. As these are implemented, the same types training and same types quality assurance has to be done on the self-service AI channels to get that confidence.
Terry Gerton One of the other points that you’ve made is that rip and replace tech models are not functional in this kind of AI world where there’s so much change and so quickly, but also that agencies may put themselves in a position where they set up significant technical debt. Walk us through how they can think about AI deployment and customer service to avoid both of those problems.
Kevin McCreight So as an architect, as a solution architect, you talked about rip and replace. It can be very expensive, especially, we’re focusing on the IRS. They have a large on-premise Cisco-based phone system. Because of the large volume of calls they have, replacing that in the near term is not an option. So the AI solution that the IRS is going to implement need to fit in well and adapt for that. So that gets to the point of, I need to find the solutions that are gonna integrate with, they’re using the Cisco UCCE solution, what works well for that? What integrates well for that? And that’s what they need to look to do. And that’s, what any agency would need to work to do. IRS, you know, another component of the taxpayer service, not necessarily self-service, but AI, is the IRS is using Salesforce to modernize their agent desktop. What tools does Salesforce bring as opposed to bringing in a third party AI? And how can those work in conjunction with their phone system and things of that nature? The technical debt piece that’s always hard because the second that you implement a solution, you’re building the technical debt. When I was at the IRS and we started to move towards the AI solutions, modernizing self-service tools, we went with at the time the Nuance product, now Microsoft product and built it out there. But for various reasons, the IRS is needing to shift and move to another one. Even though we believe we’re on the right path, we were building technical debt, and now for various financial and technology reasons, are pivoting. In the future, they may pivot, the IRS or any agency may pivot to a contact center as a solution, a cloud-based solution. There’s a lot of technical debt that will need to be undone. So as they make the business and technology decisions on the technical debt side, that’s always part of the equation and consideration. Yeah, so avoiding it can be very challenging. You always want to minimize it and think ahead, what is the best solution and how can I move things or how can I evolve with the technology I have?
Terry Gerton It’s very important framing. And as we think one more step in AI deployment in this space, a lot of the conversation here will focus and has focused on the customer’s experience. What are you seeing in terms of the workforce’s experience as AI gets deployed in customer service functions?
Kevin McCreight You know, the biggest impact for AI in the employee’s experience or the frontline employee experience is the commonly referred to as agent assist. So as when a call or a chat is transferred to the agent, being able to transfer that self-service communication so the agent has that by their side and it can interpret and recommend to the agent what next steps may be or in real time as the agent is there, as the agents communicating with the citizen taxpayer constituent, you know, what is being said, what things could be recommended? That’s a different, you know, model as opposed to having an agent fully trained where they can use their experience and their procedures to look things up. Somebody is recommending in real time, no, you should be going to this procedure. You should be starting to go through this process to answer the question or address it, or even pop up the information for the answer. It’s a bit of a training thing. The best part, or I would say the utopia nirvana of that is, if the AI tools are implemented properly, the agents don’t require as much experience and training to provide that same level of service.
Terry Gerton Let me just follow that thread. If the AI tools are implemented properly, what should a taxpayer or a customer of any other agency notice first?
Kevin McCreight They should notice getting to the right answer quicker. I mean, I think that’s the end of the game, the right and faster as opposed to having to do extensive research. Be less hold time to be perfectly honest throughout the experience. As an agent would have to say, excuse me, I need to go research your account or I need go check with somebody else. They are often times reviewing procedures, looking at documentations, looking up an account, where AI could be bringing this all into a single viewpoint or bringing it up so that the assistor, the agent, could be providing that quicker to the taxpayer and not only providing the information, but if an update needs to be made, making that update quicker.
Terry Gerton A faster answer to yes.
Kevin McCreight Exactly.
Copyright
© 2026 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.
Click Here For The Original Source.
