the voice of customer experience technology
Front pagesponsored · AWS
CX AI2h · 13:31 BST · 1 min read

Why Resolution Is No Longer Enough for Customer Experience Leaders

Pasquale DeMaio, Vice President at Amazon Connect, explains why mediocre service quietly erodes loyalty, how AI can reveal what customers really experience, and why customer experience leaders must move beyond deflection to build more connected journeys.

Author transcript

Rob Wilkinson: Hello and welcome. I'm Rob Wilkinson at CX Today, and today we're taking a closer look at why resolving a customer issue isn't always enough to earn their loyalty. If you're challenged with improving customer retention, advocacy and brand preference, or if you're exploring how AI can help create more relevant and connected customer experiences, stay with us. Today I'm joined by Pasquale DeMaio, Vice President at Amazon Connect, who has a clear view of how organisations are rethinking CX right now – not simply as a cost to manage, but as one of the ways a business can build stronger relationships and long-term growth with their customers. Welcome, Pasquale. Thank you so much for joining us. Pasquale DeMaio: Thanks. It's so good to be here. I really appreciate the time today. Rob Wilkinson: Let's set the scene for our audience a little bit, Pasquale. Your recent blog made a fairly uncomfortable point: failure might not be the biggest reason why customers leave – mediocrity might be. For our audience, sum that up. What do you mean by that, and why is it so important right now? Pasquale DeMaio: Well, the thing about customer service is that for years, people were so focused on the idea that they wanted to get rid of their customers. Deflection, containment – things that I don't think most customers would say sound like the kind of experience they want to have. Certainly I don't look to be deflected or contained. It almost became a mantra within the industry. The problem with that is that nowadays, customers have so many more options to get support when they have a challenge, when they need help or when they want to get something done. And more and more, if you're deflecting someone, you have to ask yourself: who am I deflecting them to? Am I deflecting them to ChatGPT? Am I deflecting them to my competitors? Because I bet they would love to welcome them with open arms. If that's the case, your customer's loyalty – their feeling of mutual respect – is going to erode quietly. They're just slowly going to stop coming to you. And before you know it, does it really matter who you are in the equation? Does your brand matter anymore? The answer is no, if your customer now thinks ChatGPT is the answer and you're just one in a list of options. So I think you have to be very careful if you think the legacy mental model of deflection and containment is where you want to be. Instead, you need to think about engagement and delight, and really meeting customers where they are. Those are words I've heard people say historically, but I think now they really ring true for the first time. Rob Wilkinson: It's been such a long time that we've been using the phrases "deflection" and "containment", and no one ever thought at any point that those were negative terms. It's strange in hindsight – you look at that and think, why didn't we clock onto that before now? I guess many organisations can easily show resolved queries. They can show reduced wait times. They can demonstrate they've moved a customer through a process, so to speak. But that picture seems to fall short nowadays. Where is that picture failing when leaders are trying to understand whether customers actually want to come back? Pasquale DeMaio: Exactly. If you think about what the experience was, often in the best case it was mediocrity – and mediocrity just isn't good enough anymore. When you look at something and say, "Yes, it shows up as resolved," the customer may have just given up and decided it's not worth it. When their choices were limited, or switching was hard and switching costs were high, I think that was potentially a strategy. It's not one I like, and I don't think it's one the most customer-obsessed service companies use, but it's one that could work. And yes, reducing wait time is certainly a good thing. I don't think anyone prefers to wait. But now we can do things that are very different in terms of engaging people early on, understanding their problems, and then giving them something that fundamentally feels different because it understands them. When I think about the best experiences I've had, it was a combination of the person – or the technology, depending on how I'm interacting with a brand – really deeply understanding what I'm trying to do and who I am, and then showing a real effort to meet my needs and helping me achieve that. I think most consumers are pretty reasonable. They don't expect to get a refund for something they bought, used and was great. They're not looking to scam. We do have some of those challenges, but really, most customers just want to be treated well by a brand they feel understands and trusts them. So the opportunity here is huge for leaders to rethink how they're engaging with these customers. And I think that needs to go not just to the people in customer service, but across the entire swathe of the business – marketing, the CMO, the CIO, the CEO. Everyone needs to be thinking about their customer now as truly a customer they need to delight and engage. Rob Wilkinson: It's so true, because it's often looked at through a customer service lens rather than the full customer experience and the whole journey. We don't always look at that, and it's so important that we do more of it as an industry – not just with technology – because it means the customer is at the heart of everything you do and is considered at every stage of the process when you're building things. Something else you've talked about that I liked was the "quiet exit". This is where a customer doesn't necessarily complain or escalate, but they definitely move on, because the experience was adequate but forgettable. For our audience's benefit, what are the signs leaders should be watching for before that indifference turns into actual churn? Pasquale DeMaio: I think every touchpoint with a customer is now an opportunity both to understand them and gather data, and to impress upon them the value of your product, your company and your service. So the number one thing is to look at whether your customer is continuing to engage with you – how they're doing it and where they're doing it. Historically, a lot of folks have looked at NPS scores or CSAT, where they've run a survey. And the only people who ever respond to a survey are either incredibly happy or, more likely, incredibly unhappy. So the data points you're getting show where you're really screwing up, or where you happened to do something very good. I don't think you should ignore those – when you're really screwing up, please stop. But instead, look at every touchpoint and start to measure the repeat engagement of the customer. For the first time in a long time, we can actually start to do that in more meaningful ways than ever. But we have to bring together the different touchpoints we have with customers into a single system, a single set of data, and then start reasoning on that in a much more powerful way. The wonderful thing is we now have the capabilities to do that with AI – in agentic ways that feel natural and good for the customer to interact with when AI is the right choice, and to support humans when humans are the right choice – and we can do it across every touchpoint. That's a fundamental change. Rob Wilkinson: Your point about surveys and NPS is interesting, because it's not just whether the responses are skewed towards the really happy or really unhappy. It's also coverage – is it genuinely a representative sample? Typical approaches like that cover less than 5% of conversations, whereas nowadays it's a whole different ball game. So if an organisation wants customer experience to be a growth engine rather than just a service cost to control, what happens in the C-suite? How are the CCO, the CMO or the CRO changing their priorities? Are you seeing that? Pasquale DeMaio: Yes – you brought up a bunch of really interesting points there, and I've definitely seen the mental model changing in the C-suite. A lot of things over history have caused inflection points that weren't necessarily intentional. There's an old saying: never let a good crisis go to waste. Over time, we saw the cloud fundamentally change the way people interacted with brands and businesses. We saw COVID have some pretty negative impacts, but one thing it did do was force a lot of organisations to rethink how they provided service quickly, without customers having to go to a physical location – to engage with customers more directly and to really update a lot of their technology. While I would never wish for that, the silver lining is that we now have much better capabilities to engage with people. We've seen another whole sweep of this with a new set of AI capabilities that now allow folks to understand every interaction – not just a small set, not the 10% or 2% people were always doing, but every interaction – and to do it in real time, across massive sets of information coming in from every customer touchpoint. You asked specifically how that changes the way the CEO, the CCO, the CMO and the CRO think about it. The answer is that the lights are turning on. They're thinking: if this is possible and my competitors are going to do it, I need to think about customer engagement, customer service and my customer fundamentally differently. That gives us the cover to have those conversations at levels where they were very difficult before. There have always been some forward-thinking, customer-obsessed companies that got this right off the bat. I'm proud to work for one I believe is, and I think a lot of people pick their companies because of that. But the idea that you can get by without being customer-obsessed is rapidly changing. As that percolates up and across every vertical and every industry in the world, people are having a different conversation at the C-level. That's really valuable for someone like me, who's just dying to help people understand their customers and do better by them. Rob Wilkinson: It feels like we're almost at a tipping point, because it's really exciting that those conversations are finally happening. When you can put numbers on things – which you can now – that always helps in the boardroom. It means everyone's going to have to catch up and get in the game, because otherwise their competition is going to take customers away from them. I really appreciate the context. I think we've got a really good understanding of the business and customer angle, but we can't have this conversation without talking about the technology, because it's so important and these are such exciting times in terms of capabilities. It's definitely a new era of technology, and it's moving at a rapid pace. So let's get a bit more practical. If an organisation wants to make its customers feel known at scale – which is another challenge – how does it do that without creating another layer of complexity? Because we're very good at making things difficult and building big stacks of stuff that fall over more often than not – or worse, ending up with a very polite automated dead end for customers. Pasquale DeMaio: Customers get me excited to come to work, and technology keeps me energised while I'm there. I can't imagine where we'll be two years from now, given how far we've come just in the last year. We've done a bunch of launches in Amazon Connect Customer very recently that really empower people to have the kind of conversations they want with their customers. If you don't mind, let me give you some examples. I talked about analysing every conversation, which I think people should be doing all the time. That's been possible in Amazon Connect for a number of years now, and we had a number of customers who turned it on for every single interaction because they wanted to know what was happening at every touchpoint within the contact centre. That was a very powerful but narrow view. Now we fully offer the ability to bring data from across all these different touchpoints into one customer profile – to understand all those engagement points and see the behaviours of that customer over time. Are they still buying from me? Are they still talking to me? Are they coming to my website? Do they pick up the phone when someone calls them? All the things that demonstrate engagement. And when you start to think about all the interactions they have, you think very differently about what any given interaction should be. One of the big things we've done in Amazon Connect Customer is make it very easy to understand why someone's calling or reaching out – or even to reach out to them proactively before they do. Airlines love Amazon Connect Customer – I think the top five airlines by passenger volume all use it now, which is awesome. Here's an example. Let's say a flight gets cancelled, which does sometimes happen. The right answer, hopefully, is to immediately rebook the person and send them a message telling them so. But that might not be the right thing, or the person might be very stressed and not want to leave it to fate. So let's say they call in to say their flight's been cancelled. Instead of saying, "How can I help you?", before they even say anything, you say: "I see your flight was cancelled, and I've booked you on the next one tonight. Is that what you wanted, or do you need another option?" And the person can say, "Yes, that's great." Now they're comfortable that it's really happened. That's a good outcome – they didn't have to sit through several interactions to get to the answer they wanted. But a lot of the time that's not okay, because they booked that flight for a reason, and if it's not at the same time, maybe it doesn't work for them. So they say, "No, I want to rebook on a different flight." Now you're on the phone. You want the urgency, response time and nuance of a voice conversation, which can be incredibly powerful. But picking a flight – trying to remember the first option by the time you're on the third, with different stopovers and all these pieces – is incredibly painful. So what we've built into Amazon Connect Customer is the ability, with agentic AI, to provide an incredibly powerful experience: understanding what the customer is asking and what they might want, and at the same time literally showing them the flight options within the app. They can say, "Actually, I want the third one," having that conversation in real time and making selections while getting the full multimodal experience. Think about what a game changer that is. You went from being incredibly stressed, needing to find the right flight, thinking, "I can go on the web page, but I'm not sure it's built to handle cancelled flights – or I can call, but then I can't see what's going on." We've taken all that pain away and replaced it with something you look at and say, "That is delightful." This is live today with our technology at United Airlines. It's a really great experience, and it's always getting better because they continue to innovate and extend it to more areas. It's awesome compared with what you probably experienced before when a flight was cancelled, and I love that we're bringing these kinds of things to customers. I could give you ten more examples, but I don't want to monopolise the entire conversation – you've got me excited. Rob Wilkinson: That's a great example – I love it. Dare I say it, in decades of CX, this has almost been the holy grail: how do you connect all the information you need to actually understand the customer's journey? Going back to what we've already talked about, it's not just sample sizes – it's that we'd only see that one interaction. Very rarely could we link up information about the experiences our customers have, recognise it, and deliver real-time experiences tailored to what's actually going on, rather than starting off by asking loads of questions – which is the most infuriating thing as a customer. Not feeling known, not feeling seen. So these really are exciting times. Because AI gives organisations more ways to understand that context and respond in a timely, relevant way, where do you think it can make the biggest difference in helping customers feel known – and where should leaders be cautious? In your role you have a very unique vantage point across lots of organisations. What would your tips for caution be? Pasquale DeMaio: What I see is that most of my customers are excited and love the fact that, for example, you can go 100% agentic with Amazon Connect Customer, or 100% human. That gives them a lot of flexibility. I don't think anyone's going to do either of those things – everybody wants the best of all worlds. The way you do that is by delivering deterministic workflows and AI that support both. That makes agentic experiences incredibly reliable: if you're doing a balance transfer, it's done exactly the same way every time, so you're 100% sure it's what the customer wanted and you understood it, with no risk of hallucinations. And on the back end, when a human being is the best choice – which is still often the case; there are plenty of times I want to talk to a human, and plenty of times I don't – you understand when that is, and bring those humans in with technology that empowers them to focus on the customer on the other end of the line. That's incredibly freeing and empowering for them, and they love it. We want to make those people superhuman at what humans are already good at: empathy, engagement, and understanding really complex, weird, unusual problems. Those two things together are what's getting customers comfortable with this, because as excited as they are, they're also nervous, or they've had failed experiences. The number of customers who come to me and say, "We tried a pilot with a startup that promised this, this and this, and it just didn't deliver" – my Rolodex, so to speak, is full of those. So we come back and say: your challenge isn't that agentic AI can't bring tremendous value – it's that it doesn't fix everything. But where it does fix things, it's incredible. Take a customer who says, "I want to change my address so I can do a balance transfer, because I moved last month and I can't find it on the web page." A legacy voice bot or chatbot couldn't help with that, but agentic AI can very quickly say, "It sounds like you need to change your address and do a balance transfer – is that right?" Then we can step them through it, again using deterministic AI to make that experience safe and great. We're also building Amazon Connect Customer so that when an experience moves between the different things that power it, there's no change in the outward impact on the end customer – what it sounds like, the voice they're hearing, the way it talks, the smoothness of the experience – because we're taking the best of all these worlds together. You can even bring third-party AI into it. We're extending the A2A specification, which allows agents to talk to each other, so a company can build its own AI and plug it directly in alongside the AI built into Amazon Connect Customer – or use a third party if they find one that's better than what we offer. My goal is certainly to make ours the best in the world for every solution, but I also know there'll be times when customers build their own or find a competitor they like better. We want that to be an incredible experience, not something where two systems are fighting. I see so much in the world where we're getting back to the bad old days – a Nuance IVR fighting with a homegrown IVR, going back and forth with two voices, transferring you, and you answering the same question three or four times. That's not the way we need to go. Instead, we need a fully integrated experience that takes you from the very beginning of a touchpoint – even if that's a chat or a web page interaction – potentially through multiple different agents to a human agent, and have it be seamless. We're taking care to get that right from day one, rather than trying to bolt things together, which I think is another huge part of how we get this right with the businesses that want to support their end customers. Rob Wilkinson: That point you've just made is amazing. We're in a situation where vendors aren't putting a walled, ring-fenced approach around their tech – because historically that's been a real problem, getting your data out of your platforms as an end-user customer. That limited approach also prevented organisations from doing what they can now do: choosing the level of human interaction, or the level of agentic service, they provide. So it's exciting times for customers, but also for the people on the front line who are getting support in roles that are evolving too. On that point, we know AWS is talking about AI-native and agentic customer experiences. How does that change the way organisations design, test and improve, compared with legacy approaches in CX? Pasquale DeMaio: I think the best companies are doing what you're describing. My concern is that there's a lot of impetus in the broader industry pushing people down some of these legacy ideas. For example, I see outcome-based pricing where the outcome, again, is deflection or containment. And I think: it's good if you solve the problem. But can you really trust that the outcome was the right one for the customer? That's always a little more interesting. And when you couldn't solve the problem, was the experience now twice as bad for the customer? Are we heading backwards into the bad old days? I think in some cases the answer is yes. But we have the opportunity right now not to do that. I think the most forward-thinking companies looking to help with AI technology – and certainly the businesses who really care about their end customers – are going to embrace the idea that we need an end-to-end approach, and I'm seeing people get very excited about that. There's been a lot of push around this. I used to start every talk by asking people to raise their hand if they were there because their boss had told them they had to do something with agentic AI by the end of the year. Then I'd say, "Keep your hand up if you think your boss knows what agentic AI is." Everyone would put their hand down and start laughing. It got us off to a good start by recognising that the conversation had to go beyond just ticking a box, to solving real customer problems in a way that delights, not in a way that deflects. So I tell my customers: deflection is dead, and if you continue down that path, your business may be as well. This is a moment in time where we have the opportunity, and we're not going to pass it up. With Amazon Connect, we're taking our customers by the hand and really partnering with them to deliver great experiences. And with the way we price Connect, you don't pay for it unless you use it, and you don't use it unless you get value – so we're really incentivised towards the same outcome: delivering great results for their end customers. It gets me so excited every morning to think about how we can bring these new technologies to bear on an age-old problem: how do I delight a customer to the point where they'd never want to leave? Rob Wilkinson: The possibilities are amazing when you think about that. There's so much opportunity for organisations to get it right now – and it really does mean that if you don't, you're going to get left behind. I could talk to you all day about this, and I think our audience is going to get loads of value from this conversation. We're running out of time, but before we wrap up: if you could offer one practical piece of advice to an exec team this quarter – someone watching this – what should they do first to move away from adequate service towards experiences that actually matter, and that customers remember for the right reasons? And how do they make sure they don't fall into the traps we've talked about? Pasquale DeMaio: I'd say that for the first time, I think you can get to the north star you want. So quickly paint the north star of what great looks like for your customers, then move very quickly. With our new agentic customer experience designer, built into Amazon Connect Customer, you can literally take control and deliver an agentic experience in hours. You can have something incredibly powerful in days rather than months or years, and you're in control of iterating on it. So there's no danger in going fast. You can run experiments. You can roll out to small percentages. You can pick certain customers and test it. So let's go fast together on this. Don't wait. Don't get left behind. And know that you're not going through a one-way door you can't come back through. We can help you do this in a way that enhances every part of the customer experience you have today, and moves you forward at the speed you want. So go as quickly as you're comfortable with, knowing you're not going to be forced down a path. It's not like jumping off the side of a building – instead, you can take a nice relaxed slide down and enjoy the view as you go, knowing you're getting better every day. So let's do it together. Let's move quickly. Rob Wilkinson: Just before we wrap up, Pasquale – for anyone watching who wants to learn more about Amazon Connect Customer's approach, what's the best way to do that? Pasquale DeMaio: We'd love for you to come to the AWS website and search for Amazon Connect Customer, and you'll be able to dive deep there. I won't read you a blog URL, but look for our blog posts on the new agentic CX designer – that'll give you some really amazing information. You'll be blown away by how fast you can move without having to be an engineer. Business users can use this and get incredible outcomes, so I think that would be a great place to start. Rob Wilkinson: Amazing. And of course, you can discover more articles, resources and videos just like this one at cxtoday.com. But for now, that's all the time we've got. Thanks again to Pasquale DeMaio for joining us. I've been Rob Wilkinson – thank you so much for watching.

Is resolving a customer issue still enough to keep people loyal?

In this CX Today interview, Rob Wilkinson speaks with Pasquale DeMaio, Vice President at Amazon Connect, about why mediocre customer service may now be more dangerous than outright failure. Customers often leave quietly, not because one interaction went badly, but because the experience felt forgettable, disconnected, or transactional.

Pasquale explains why legacy ideas like deflection and containment are becoming risky, especially as customers gain more options through AI tools, competitors, and self-service channels. Instead, he argues that organizations need to rethink customer experience as a growth engine, using AI to understand every interaction, connect customer data, and support both agentic and human-led experiences.

The conversation explores how leaders can measure customer engagement more effectively, why surveys and resolved-ticket metrics only tell part of the story, and how Amazon Connect is helping brands create more seamless, proactive, and personalized experiences at scale.

See Pasquale's blog here.
For more Customer Experience tech news visit CX Today.

The discussion0 takes · attributed & checked

Does this reflect your experience?

opening the room…
Read nextordered by techtelligence · every pick explained
same beat · CX AI

How Customer Simulations Turn Research Into Valuable Insight

37m
same beat · CX AIIs Contact Center AI for Augmentation Missing the Opportunity?13 Oct 2026same beat · CX AI3 Consumer Expectations Defining AI in CX This Week13 Oct 2026