
Retail Untangled
Business hacks and retail insights you won’t find anywhere else. Stories from the coalface. Real time innovation and solutions. Brought to you by the team at Inside Retail.
Retail Untangled
Episode 29: The AI revolution is just beginning: How Agentic AI will transform CX
In this episode, Amie sits down with Isabelle Zdatny, head of thought leadership at Qualtrics XM Institute to discuss how AI has the power to unlock hundreds of billions of dollars by transforming how retailers understand and serve their customers.
Amie:
Welcome to Retail Untangled, my name is Amie Larter and this is the podcast where we speak to retail industry experts and find out business hacks that help them succeed. You won’t find these gems anywhere else, and we have some superb stories from the coalface as well as helicopter insights from retail industry leaders.
Now, just for a moment, imagine unlocking hundreds of billions of dollars by transforming how you understand and serve your customers. Now just for a moment, imagine unlocking hundreds of billions of dollars by transforming how you understand and serve your customers. According to today’s guest, that's the power of AI in customer experience. In this episode, I am chatting to Isabelle Zdatny, Head of Thought Leadership with the Qualtrics XM Institute about her recent report that really supports retailers that want to turn AI in CX into a competitive advantage, rather than a series of pilots. It’s a framework, a practical playbook of sorts and we’ve gone straight to the source to try and get as much practical information as possible.
Welcome, Isabelle.
Isabelle:
Thank you so much for having me. Happy to be here.
Amie:
Thank you. So our discussion today is going to center on the transformative power of AI in the customer experience. It's something that we've spoken about. We've spoken about a lot. I was just recently at Shop Talk, and I think that this was a word at the top of everyone's minds. But I think to begin today, your recent report highlights the increasing inadequacy of traditional CX approaches, particularly in the context of evolving customer experience expectations and the rapid technology disruption. Can you please elaborate on the most critical challenges confronting retail leaders in the pursuit of exceptional customer experiences right now?
Isabelle:
Absolutely. And I would say I think what we're seeing is a collision here between rapidly changing external factors and these internal limitations around current customer experience capabilities. So as you alluded to in the external environment, we are seeing three, I would say, really critical shifts. So first of all, customer expectations are continuously rising. This was one of our key findings in our 2025 Consumer Trends Report.
Consumers today are comparing the experiences that they're having with one organisation to the experiences that they're having with best in class organisations, regardless of the specific industry that they're in. Second is, I think, following on that, customer loyalty is actually in decline. So our research shows that about 65% of consumers cut their spending after a bad experience with a department store or online retailer.
So, they are increasingly intolerant of friction in their experiences and it is easier than ever for them to switch between brands. And then the third big one I would say, again, is these massive technological disruptions that organisations need to navigate today where it is constantly changing how you are discovering and shopping with retailers. And then I think those external pressures are exposing some limitations in organisations internal customer experience capabilities. And so for this report, a framework you'll hear me talk about again and again are these three activities of customer experience activities.
So first one is customer intelligence, how organisations are capturing customer data and feedback and then translating that into insights. And we see a pretty serious limitation here right now with declining customer response rates. So companies are having a harder time understanding customers' experiences. There's lags in this data, right? You send periodic surveys or doing manual analysis. That data often doesn't come in until too late. And you end up with reactive insights.
And then even when you are getting good data, it is siloed across different departments and teams, which makes it really hard to come up with a holistic understanding of customers' experiences. So that's kind of the first category.
Second one is around experience, design, and delivery. So how organisations are translating those customer insights into meaningful and personalised interactions. So as far as limitations go today, we see a lot of organisations collecting random data that they are not then able to systematically turn into better experiences for their customers.
And then the third bucket is huge, which is around operational performance. So how organisations are executing consistently great experiences while optimising their limited resources. And this tough one here is that operational complexity has just exploded in recent years, right? So retailers today have to coordinate experiences across different channels. And they're often required to use legacy systems that were designed for these single channel operations. And so that's often also getting in the way of delivering exceptional experiences. That's kind of setting up the discussion for why AI for CX is so transformative.
And the final point I would make here is that a lot of executives today actually recognise the state of affairs at 77% of executives consider CX to be either a significant or critical priority for their organisation. They get that experience as becoming a key competitive differentiator in this environment. And 72% expect AI to fundamentally transform their organisation's approach to customer experience over the next three years. So lots of limitations and AI is going to help us overcome them, hopefully.
Amie:
Yeah, I mean with those limitations comes opportunity and I think the thing that I love about this report is that it really quantifies the significant financial opportunity presented by AI within business generally, but if we look at the retail sector specifically, it projects that AI enabled CX could generate an estimated $440 billion in annual EBITDA for consumer facing businesses and that consumer retail and retail banking each stand to realise $ 100 billion in EBITDA through things like hyper-personalised marketing, predictive analytics, and the streamlined operations.
So I'm keen to understand, and I'm sure everyone that's listening is as well, now that you've identified the productivity enhancements, growth initiatives, and process improvement as sort of those three key pathways, from a retailer's perspective, which of these pathways offer the greatest potential for optimising return on investment? is this sort of like more of a combined, you need to use each piece strategy.
Isabelle:
Yes, I think something you will hear me say a lot over the course of this conversation is that you need to start with a holistic strategy of what are you trying to achieve? What are the business outcomes you are looking to achieve? What are your brand objectives? And then how can different AI solutions help you to achieve those goals? Because there are a lot of different AI use cases out there today that can create value in a lot of random places.
Amie;
And do you think, sorry, on the flip side of that, do you think that that's why we've come to this sort of, I don't know how I would explain it, you'll explain it much better than I, but I'll have a stab at it, is that we've come to this place where it feels very siloed in terms of someone will see something as, this could be a quick win, let's trial, let's implement. And it does become very siloed. I mean, and it's certainly not just in retail, I've seen it in media. Is that sort of why we're in this place of things feeling less than ideal in terms of integration.
Isabelle:
Absolutely, yeah, call it pilot purgatory and there's actually just pulling more data points here. An Axios report recently released data from their executive research and employee research and one of their findings was that executives were concerned about how siloed all of their different AI use cases and experiments are.
And what happens is if you're just doing these isolated small scale experiments in different departments and you're not connecting up those dots across the organisation, you end up running really expensive experiments that are not producing meaningful value. The real value of AI comes when it is able to learn from massive amounts of data. and if you are just keeping things siloed, are never going to hit the ultimate kind of business objectives you're looking for, you're never going to be able to meaningfully change your brand. This has to be a cross-functional top-down effort. A lot of the elements that make AI implementation really successful, like cross-functional governance structures, clear risks and ethics guidelines, an AI-ready workforce, those all have to be centralised and work across different functions.
So I think there's some trick here within the short term: selecting use cases that based on your business and brand objectives or your customer experience, maybe your CX program, goals and strategy, things that are going to be low effort and high impact. Maybe you already have vendors that have some cool AI features. If they support your product vision, go turn them on.
And then you also need to be thinking about what are those two to three like lighthouse use cases that span departments that are going to take longer time to roll out and implement, but when you do, they're gonna produce more value for the organisation.
Amie:
Yeah, and they generally sit in those three pillars I'm going to have to assume of that productivity, growth, and process improvement. Is, from your research, and I know we don't want to suggest that any one is better than the other, but is there more of a foot in the door approach with one over the other?
Isabelle:
I don't know if it's aught, but I would say is. What is happening is we see organisations early on focusing more on the AI use cases that are driving productivity gains and process improvements, because a lot of those are around using AI to accelerate your existing ways of working, just automating things, making things faster, maybe helping employees make better decisions in their role.
I think over the longer term, the place where we are actually going to see the most value is under revenue growth, right? How are we using AI to not just accelerate our existing ways of working, but unlock completely new ways of understanding and delivering value to our customers that that is ultimately going to be the competitive differentiator.
Amie:
That makes sense. So do you have any examples or what are some of the most compelling applications of AI in transforming the retail customer experience across various stages of the customer journey? And yeah, what are those best in class examples or leading organisations that you'd highlight?
Isabelle:
Yes. So just to set the scene for how we ended up creating the value model for sizing the opportunity here, which again is $100 billion in annual EBITDA in the retail industry, is we looked across a generic customer journey. So in this case for retail, things like discovery and consideration and purchase and, you know, post purchase support, and then looked at where are the biggest pain points across those journeys and where are there AI solutions that can be deployed to help close those pain points and to improve customers' experiences. And so we did look across journeys, some of the, I would say, most exciting things we're seeing is, for example, in the discovery journey.
We're increasingly seeing AI powered visual search. So helping customers find products, retailers like Wayfair, H&M allow customers to upload images to find similar items, removing some of that friction from the shopping process. We're also seeing increasingly sophisticated recommendation engines that look at not just purchase history or demographic information, but more of that psychographic information, historical browsing behavior you know, customers communication preferences, which is allowing for even more personalised recommendations. We're also again seeing hyper personalised marketing campaigns for consideration. Dynamic pricing and promotion optimisation is a big one, right? But I think for retail, especially ways that we can help create more tailored offerings and recommendations is going to help us maximise conversion while protecting our margins.
And then for me from the customer experience world, I'd say a number of the ones I'm most excited about are in the post purchase experience. So how can we do more, say predictive service models instead of waiting for customers to come report a problem. We're seeing a lot of those best in class retailers using AI to detect potential issues like delayed shipments or product defects and just proactively communicate that with the customers, maybe proactively send them a new offer, discount code, to turn those potential negative experiences into opportunities to actually build loyalty.
Amie:
Which as you've said is on the decline, so this is just so important.
Isabelle :
It's huge. And I think one of the things that I would say in retail, especially organisations do need to be careful of is when it is so easy to switch between brands, you want to be deliberately building that emotional connection between your brand and your customers. And so a lot of organisations I'm worried, are over indexing on that automation piece. And while there is value in making things easy and fast and convenient, if your only relationship and value prop to your customers is that, you can buy something from us in three clicks. If someone comes along and can buy, customers can buy from them in two clicks, they're going to leave you.
And so you also want to be using AI again for those types of personalisation pieces, helping them achieve their goals more quickly and efficiently to build those longer term emotional bonds that will have them stay with you even if you make a mistake and even if a competitor is even easier to do business with.
Amie:
I love that, so that emotional piece and I actually started talking a lot more about this now even with the conversation around the power of AI that will become a differentiator.
Isabelle:
Absolutely, and we're getting better at understanding emotions at scale, right? That's some of the things that sentiment analysis, even natural language processing on unstructured data is now allowing organisations to do that at scale. Internally, it can't just be about the data though. need to have, know, how do you want customers to be feeling? What is your brand promises? And then making sure that that's executed through human centered design processes.
Amie:
Looking forward, so agentic AI is poised to revolutionise the industry. How should retail leaders prepare for the emergence of this technology and what are the potential long-term implications for I guess business models and competitive advantage?
Isabelle:
That's a nice easy softball question. Let me tell you what the future looks like. So I think maybe just to level set here what might be helpful is the way I categorise AI is across three distinct categories. You have your analytical AI and this is the kind of portfolio of AI capabilities that allow you to process massive volumes of data so you are able to predict future events and uncover hidden patterns. So things like predictive analytics and natural language processing and sentiment analysis, right? All of those pieces that help derive better data and insights. Those have been around for about 15 years since 2012. Organisations are still not great at them.
So that's already embedded into a lot of organisations’ operations. Next one obviously is generative AI that helps you create new content and power natural language conversations with customers. That's a big one. A lot of the $860 billion opportunity across 19 industries comes from generative AI. And then the third one, the buzziest one is agentic AI.
While analytical and generative AI excel at specific tasks, they still require human intervention, right? They still require a chatbot, still require humans to prompt it. You can have predictive forecasting. You still have to go act on it. Agentic AI is able to independently orchestrate multiple capabilities across complex workflows. So rather than just assisting humans, these systems are going to be able to drive complete end-to-end processes all by themselves. And before we get there, there will need to be humans in the loop, right? This isn't going to happen overnight.
But eventually you will have these AI agents, agentic AI systems that are able to complete end-to-end workflows and processes while flexibly adapting to changing conditions and making smart decisions along the way. So they're not just following a rule-based system. They're able to be a little bit more flexible.
It's like having a fleet of these automated project managers that are able to coordinate tools and people and other systems, including other AI systems to accomplish a specific goal. So I think as you're alluding to your question, right, like as these systems become more mainstream and mature, this isn't just going to be an incremental improvement for business. This is going to fundamentally transform how organisations understand and serve their customers. It's going to drive hyper-personalisation, seamless end-to-end experiences, radical operational efficiency, lots and lots of things.
Amie:
It sounds equal parts terrifying when you explain it like that, but so, so, so practical. And Isabelle, I must congratulate you, because that may very well have been the most practical way I've ever heard that explained.
Isabelle:
There is so much confusion out there right now. And yes, I think it does sound scary. I think in the short term, there's a lot of obvious applications for it. If you look internally at your own systems, all of us spend a lot of time on administrative tasks and going through processes and workflows that we don't wanna be doing with our time, right? We are happy to offload that.
Amie:
It's so true.
Isabelle:
I think over the long term, there is a broader discussion to be had over what does this mean for the future of work, which is probably outside the scope of this conversation. But very interesting stuff. I would say as an organisation to start preparing for it today, which you should because as soon as these systems get turned on inside organisations, it's going to start creating these compounding advantages, right?
You are building deeper customer relationships, you're getting more data, smarter data, more efficient processes. So definitely worth the investment. I tend to think about preparation for this across three different areas. So first of all, you need to have a technical foundation in place, right? This technical foundation needs to be omni-channel. We can't have siloed fragmented data living in different systems across the company.
These agents are powered by data. They're going to compound the garbage in, garbage out problems with data. So starting to bring those systems together and building a healthy data diet. So we're not just overly relying on structured survey-based feedback, but able to also bring in unstructured and unsolicited data as well.
And able to build up a 360 degree view of individual customers, that when an agent is interfacing with a customer, they have the full context of that customer's history, their preferences, their real-time journey steps, everything that they need to deliver a good experience for that customer. This also includes things like structuring your knowledge bases, your standards, right? Making sure that they're querying good data. One of the problems we've seen.
Recently, is contact centers building out internal co-pilots to help support their agents. And the co-pilots are trained on data from like 2001 knowledge base articles. And so they're pulling up nonsensical and wrong information for the agents to take action on. So you have to get those cleaned up too. From an organisational perspective, you need to define your AI operating model.
So again, starting with that clear, clear vision. What are we trying to do here? What types of experiences do we want these agents to be delivering that needs to be aligned with your experience and business goals. Gotta establish robust risks and ethics guidelines. And then also start building up those cross-functional governance structures. So we're seeing a lot of companies start standing up central AI councils, kind of co-opting previous councils that will have like legal and security and IT to be the arbiters of what is an appropriate AI solution, and how we want to implement consistently across these different departments. And then the biggest one I would say is preparing your people, building that AI ready workforce, starting to get people comfortable with.
Amie:
Yeah, unpacking what sort of sounds oddly terrifying to really expose the real opportunity and it is such a great opportunity and I'm keen to understand just quickly, you know, we mentioned at the start of this that we're really at that space where we're still trialing, some people are still trialing AI, there's no real, there's not a lot of strategy behind it for many organisations, yet we're rapidly approaching an environment where you need to be ready so that you can really make the most and harness this type of technology.
There seems to be that gap. And so if you're not there yet, you will be left behind. And the people that are really going to power forward in terms of hyper-personalisation, being able to really cater and win in terms of customer acquisition. For those that are looking at that and going, okay, you know, where do I start now? Or, you know, am I too far behind? What's the message there?
Isabelle:
I think, just you said that so well, and just to add, in case people need even more convincing, one of the things we found in this report was that 90%, nearly 90% of organisations have some AI initiatives underway across their organisations. Most companies are doing this. However, these activities remain limited and uncoordinated. So only 12% of executives in our executive study said that they had an organisation-wide AI strategy in place with coordinated ownership. Of that 12%, they were 2.4 times more, sorry, 2.3 times more likely to report market share gains compared to the rest of the group that was not taking this type of systematic approach to AI.
And as you said, it does create these compounding advantages: every customer interaction makes these systems smarter, every insight builds more value, every experience improvement deepens those relationships. So for getting started, again, I think it comes down to: start by defining your vision. You need to have a clear unified approach to make trade-off decisions. Again, there's a lot of different AI solutions out there. I'm sure everyone's doors are getting batted down by vendors saying, hey, we have these cool new AI offerings for you, you need a set of prioritisation criteria to help you identify where you should be investing your time and effort. And that starts with having a clear vision and strategy in place.
Amie:
So from that vision point, I mean, and obviously we've got to the point where we can see barriers to implementation or that strategy piece. And that's where I think, I've got, I think that's where we need to be, we all need to be. We need to be moving beyond just testing these things intermittently for whatever small wins we might get and having that vision piece. What's holding us back? Is it skills, not having the right skill set to be able to create a vision? Why aren't we getting to the vision piece?
Isabelle:
Yeah, I think it's, people are nervous about if AI goes wrong, it can go very wrong. It can cause irreparable damage to your brand. So you exactly put your finger on one of, think the real interesting disconnects in this study that so many executives understand the transformative potential of AI. Like, almost 70% think that AI is going to completely transform their industry within the next three years but only 15% of them aspire to be at the forefront of this AI driven business transformation.
I think there's this wait and see approach because it can go so wrong because there are not existing playbooks and it's scary to be the organisation that's defining those playbooks. One bad instance of, you know, an AI bot going rogue or whatever it is can do some real lasting damage to your brand. So there is this tension between moving decisively to capitalise on this compounding value while also being responsible and ensuring that you have the proper guardrails in place that you are not going to do irreparable harm to your business. So I think there's a lot of fear out there, honestly.
Amie:
Yeah, and I suppose that has always been to innovation more broadly when we think about it over time. It's easier to let someone else invent and then follow suit. It's cheaper, most likely. However, I think in this case, the big wins that we're really seeing is that that deeper personalisation and relationship with customers is really going to pay off in the long run which is obviously really cool and going to be a standout for retailers.
Off the back of that is when it comes to cultivating that culture of innovation and collaboration and change management, what's your advice to retailers right now?
Isabelle:
Oh man, that's another big picture question.
Amie:
Surprise!
Isabelle:
I know. Yes, so I would say, again, for our work, we tend to think that exceptional customer experiences require three key pillars. So you have the technology pillar, right? You need AI, you need the technology infrastructure in place that's going to allow you to scale and enable excellent experiences. You need the culture in place. So we look at four specific attributes that lead to a customer-centric culture. Again, one of them, feel like I've been harping on this a lot, is being purpose-led, starting with that clear vision that also helps people across the organisation know and be able to make decisions in their role without a command and control structure of like, is a AI solution that we should turn on or this is an experience we should go deliver because we understand the executive vision that we should be aligning towards.
Another attribute is human centric, so being empathetic, talking about emotions. I know it feels a little squishy and uncomfortable for most organisations. Organisations need to focus on being change-minded, so seeing failure as okay and learning experience, so being okay with a lot of different pilots and activities in different parts of the organisation, key for an innovative culture. And then also being change-minded, making sure that from the top, your modeling behaviors, that you are not making decisions based on gut instinct or how you think people behave or will react, but on actual data.
And then the third pillar of a strong customer experience program, strong customer experience operations is that competencies piece is building these skills and actions that you need internally to consistently deliver standout customer experiences. I would say on that one, you probably need a customer experience centre of excellence, someone who has authority across the organisation, not just buried in a different department or function like digital or contact centre to align and drive consistent customer experiences across the organisation. But none of this is going to happen overnight.
So I think getting really clear on where you are today, where you want your experiences to be, especially in an AI-filled world, and then how you need to move the culture, build the skills and competencies, invest in the right technologies to help you get there. It's not going to happen overnight.