Harnessing AI to orchestrate customer experiences at scale - Interviews with Luc Dammann, Nina Caruso & Vivek Pandya of Adobe

28:09What does that mean? It means there has to be someone who is responsible for the customer data platform. It means there has to be some agreed-upon ways of working. There should be an executive stakeholder that is helping with budget, but also to unblock any kind of organizational. Alignment feels like a thing that kind of comes up a lot. Alignment is big. But and also in terms of alignment is having a North Star, having some type of vision that everybody who's working in this organized group is rallying towards. I mean, thank you for saying that because that's something I, it's one of my drums I get out from time to time.

28:51Well, actually all the time is sort of bang it. And because I always think that it's all good having all these fancy tools and toys and all these different things. but if you don't have a clear idea of what you want to do and why and where you're going and you don't have everybody aligned around that then it doesn't really drop out in terms of what's needed by who and what and when in order to enable that because otherwise it just feels like everybody's doing their thing and it all feels a bit and people wonder why oh it all feels a bit messy yeah you're like going hmm it's it's not it's not rocket science yeah as it were.

29:27No, yeah, these are really like pretty straightforward, obvious things, but it can be hard to get people on the same page and also just devote, you know, the kind of time to think through this. And I think that's one of the big things, it seems to me as well. I'd be interested to hear your perspective on that, because I think prior to this role, you also led a lot of these technical kind of implementations and things. So you'll have been in the weeds and you've seen all these different things that oftentimes taking the time to achieve to understand different people's perspective and then achieve alignment and things can feel a lot like you're talking about talking about it in many ways and it sort of runs almost counter intuitive to the sort of this modern organizational operating rhythm in many ways people are like oh no we have to be going off and doing making things and doing things but it's time well spent yeah Yes.

30:26We make the argument that spending the time on this pays off. Right. Pays off in terms of efficiency gained, can help pay off in terms of getting to some of the more bottom line key performance indicators or even just general return on investments of the technology. And Adam covered that in the session of what were some of how did they get to their goals? And in part, they were able to get there because they spent the time, you know, creating a framework that worked for their organization. The other thing that you talked about in the session, I think that I wanted to pick up on, was you talked about being very clear that adopting this type of CDP platform in order to drive this, as Adobe likes to talk about, it's a personalization experience at scale, requires, and is going to require new roles and skills and responsibilities and things.

31:28And I wonder if you can speak to, tell me a little bit more about where you think that organizations need to mature into or to build capability kind of into because it's, there are some gaps in the capability gaps in this, it feels to me. So where are some things that organizations need to be thinking about, where they need to be building skills and capabilities in order to deliver that? Is that fair? You know, it's interesting. I wouldn't make the argument something that that we talked about in the session, too, is, hey, you don't need to, like, necessarily go out and hire a bunch of people to, you know, run your CDP.

32:10But I think it is important to think about as you're building some type of operating framework for a CDP to whether that's a center of excellence or more informal to think about having a diverse set of skills. That can be across multiple people or it could be within one person of people who have an understanding of the enterprise data, who know where it exists and can be able to bring it into the CDP. Important to have some people, too, who may become from more of an analytics background and can help with some of the measurement. But I think something you were hitting on in that question was what are new skills that we need to, you know, start to think about as we move into customer experience orchestration in the age of AI?

32:58I think what's going to be really important, and I think about this in my own kind of day-to-day work too, is how do we all get more comfortable of working with agents and thinking of them as co-workers in many ways? of thinking more about how do we integrate agents successfully into some of the existing workflows that we have to think through how do we give ourselves time back to move away from some of the grunt work, if you will, of those repetitive low-value tasks and give ourselves the time back to focus on more strategic endeavors. Yeah, I think that's fair. so thank you for asking kind of the question i was trying to ask in a better way than i was asking it so that's kind of very helpful but kind of just wanted to revisit some of the the air miles sort of experience and what adam was talking about because i think it's it's it's interesting to highlight i mean what were some of the tell me a little bit more unfortunately adam's not here but can you tell me a bit about adam's kind of like story and air miles because story and what because you mentioned some of the things the challenges that they had to face and particularly the stealth CEO.

34:15Yeah, we're going to remember that we can talk about the stealth thing. Tell me a bit more about what they did, the challenges they faced, kind of how they overcame them, and actually what some of the outcomes have been. Yeah. What's interesting about them is they are a very data-centric organization, and they wanted to move from, we don't want to just have all the data for the sake of having the data. We want it to, we want the data to work for us. We want to be able to activate it, to use it as a means toward an end, and that being customer engagement and growing their user base and cross-selling and upselling and things like that.

35:00What Adam will walk through is that this is something, the figuring out the operating framework for them and the people and process element of this. He'll, you know, tell you this if you were to speak with him, is that this didn't go perfectly, like, right out of the gate. There was a lot of test and learn mindset to this and needing to spend time of figuring out and spending that intentional time of, okay, we tried this. This didn't work. We need to tweak this to figure out what are the what's the right combination of processes and people that will make the CDP work for us. And he'll he would tell you as well and covered in the session that there was he talks about this as transformational fatigue.

35:53Right. Where they were you know we've been in this you know decade plus time period where digital transformation has been such a buzzword and figuring out how to adopt new technology, how to organize people around it. And I think their organization was left exhausted by some of them and wanted to be able to move beyond, you know, some of the time associated and effort associated with more formal centers of excellence. So that's how they landed on kind of this stealth COE, where they had the people in process, but it wasn't quite as formalized as it had been in the past. Some of the things they were able to achieve by finding the right model that worked for them and taking that time to tweak it and optimize it so that it worked for their organization were being able to have a higher coverage of audiences used across the organization, of having one repository where audiences are built and making sure that they're not just built for the purpose of building them, that they're actually being used and activated and decreasing time to market as well, of being able to take a request or an input from a marketing team or a paid media team and being able to much more quickly get that out.

37:23Okay. I love the idea that about transformational fatigue because I also think about there's there's other things I talk about when people launch another transformation initiative that can sometimes in people can get corporate indigestion as it were you're going I'm getting heartburn from from that sort of thing and so it's I think it's really smart to you have to understand your context yes in order to be able to to deliver something and I think the whole stealth thing kind of makes kind of like sense because you're like going nah the thing we're doing or the thing we want to do is not wrong but actually contextually it doesn't kind of like make sense to do it formally and therefore we're going to do it informally and i think that's really that's a really important kind of point and particularly for people listening and it's like going you know just formal may work in some context and then sometimes but sometimes you You need to be a bit more stealth or a bit more guerrilla in order to, because it's all about the outcomes and whatever outcomes you want to do.

38:25And you've got to figure out the best way to achieve that. Yeah, I think it's important to recognize that finding a right kind of people and process and operations fit is, it's not one size fits all. And it has to be, you know, tweaked to be successful. so you know one final question and it's something i wanted to kind of like uh to draw because you're talking about cdp it's now right now where it feels like we're in an environment where people are talking about real-time decisioning orchestration contextual intelligence all these different little things and people are like wow because that's going to be cdp or it's going to be crm or you know some people are going on it's not going to be these sort of ccasts because they've got that they're having all this real-time interactions and things and it's been like oh that feels like it's a bit like a bit of a big street fight but scuffle sort of thing kind of like people are having to figure it out i wanted to ask you just get your position on it so why do you think that i mean one do you agree with that in terms of these are the kind of the the big platforms that are emerging and two why do you think the cdp is best position to sort of to deliver on some of the promises that we're talking about.

39:43Yeah, I agree that there's a lot of technology that is playing a... Jostling for position. Yes, that there's some jostling that is going on now. It's my belief that in many cases, organizations are using multiple of those different players to get towards an overall personalization at scale and customer experience orchestration and the goals that they have for it. So for example, with our customers of real-time CDP, it's very common to bring in CRM data, right? That they have an existing CRM, they're bringing it in a subset of that data into the CDP and then using it towards activation. Where I think customer data platforms have such a critical role to play, like I mentioned at the top, is in creating this really powerful foundation of data for the purposes of activation.

40:48And that being so critical now in this moment that we find ourselves in as, you know, there's new data emerging because of AI, there's new channels that it can be used in. I think what customers sometimes think of AI is great. I can move faster. I can do more. In reality, it won't serve as any type of silver bullet or even a Band-Aid. What we find is that sometimes it can expose the cracks in the data foundation and can expose flaws in it. And we see CDPs as critical to helping customers create that solid data foundation that is harmonizing the profile across structured and unstructured data that has that governance layer so that you can really confidently activate data.

41:42There you go. Nina says it's thus. That's all I have for now. But I just want to say thank you. Congratulations on this session. Thank you for sharing some time and some of your expertise and your insight with me. That's been super cool. Thank you for having me. Appreciate it. Well, that was another great conversation and some really fantastic insights. A couple of things that stood out for me include, one, how AI exposes data flaws and that working with AI agents as co-workers is becoming a new skill that we all need to start thinking about developing. Lots to think about there. Let's keep going and let's get into my next channel.

42:16This time I speak to Vivek Pandya, who is the lead insights analyst across Adobe. We're here at Adobe Summit again. This is another featured interview with an Adobe expert, but also a friend of the podcast, Vivek Pandya, who is a lead insights analyst at Adobe. Is that right? That's right. Perfect. Welcome back. Thank you. How are you doing? I'm doing well. I'm doing well, especially given day one at Summit today. I'm holding up well. Perfect. So here's the thing that I wanted to start with. I wanted to start with that. So you and I had a chat on the podcast about a year ago. In fact, it's almost exactly a year ago.

42:58And there was this, all of this, these kind of trends, all these insights that's coming out about consumer behavior, particularly around AEO, GEO, call it what you will. and you were sort of saying, oh, this is some emergent stuff and didn't quite know if any of it was going to stick or not. A year on, what's changed? What's sticking? Where are we at? I think we're still in a pretty aggressive growth phase with this type of traffic. So before we go into it, I tell you what, can I ask you to frame kind of what's actually happening? Because then we can think about sticking. Definitely, definitely.

43:33So what we're seeing is consumers leaning on generative AI platforms to get the sentence of what they want, what's available, do shopping lists, all these types of behaviors. And we previously saw this traffic scaling and meeting and dropping on retail sites last year, and that continues to accelerate. So just here in Q1 2026, we see traffic from generative AI platforms up 393 % year on year. So there's very few things that are growing quite like that. Yeah. So it's, it has huge implications for the state of comfortability that consumers have with using these platforms for all different use cases across industries, but especially retail, where they want to understand what's available, where they can find cheaper versions of goods, where they can find certain items for specific people and kind of get it comfortable when you shop for other people.

44:34So that's been very exciting to see. I think compared to a year ago when we first talked, generative AI platform traffic was converting, kind of, it was growing in terms of conversion, but it was still underperforming general traffic to sites by about 38%. Okay. now here we are a year later this traffic is over performing non-ai source traffic by over 43 percent so you've gotten a situation where consumers are very comfortable you know having that dialogue with the llm agent seeing what's available to them and now they're just very comfortable with that journey and buying on site. So it's gone from being sort of a kind of a cursory exploring pages, you know, just a journey that consumers are, you know, kind of testing and experimenting with to an everyday operationalized part of their life where they're leaning on these platforms for ideas.

45:38They're thinking about what's available. And ultimately we're in a state now where about 40 % of consumers are using AI in some level of their shopping journey. So this has huge implications for how brands think about, you know, the sort of long-term trajectory of this. So let's unpack that a little bit because it makes me think that either, excuse my language, either e-commerce websites are rubbish and it's hard to navigate or to and or consumers are lazy and is it the best i would say consumers are looking for ease and efficiency okay that's a more that's a more pleasant way of putting it it's it's really you know if you look at what people are dealing with in the everyday world there's a lot going on and they want shopping to be very straightforward they want the peace of mind that their choice will be you know well vetted and it will be sort of confirmed to them in the fact in the sense that they'll be happy and satisfied with their order and they won't want to return it so that's where the llm is really coming through because we're seeing even through survey research we obviously look at a lot of our data through the overall adobe analytics data set we're able to get some really great dynamic journey data from there but we also do survey research and we do see instances where consumers aren't returning as much if they go through this journey of, you know, using AI through their shopping path.

47:10So it's very indicative of a consumer who has had to deal with, you know, the everyday day-to-day purchasing in shops. And then you go on a website, there's a lot of, you know, options, you might have things popping out at you. And the AI experience synthesizes all this and it distills it down and the successful outcomes consumers are having with it and the fact that they're trusting these results is resulting in them using it more that's our explosive growth and it's driving stronger purchase intent and that's our conversion rate over performance so this is a real sort of paradigm shift for these brands to think about because how do you ensure that you're getting the right visibility in in this type of environment and you also have to make sure that your site is ready to absorb traffic from almost a pre-qualified generative AI lead.

48:06So you have that back and forth conversation, you're understanding, you're getting a rich sense of everything that's available. And then once you get on the site, it's, you're again, seeing this traffic is very research focused, they're strategic and conscientious, they're spending more time on site. So that's resulting in them being very comfortable purchasing and so that's a pretty impressive metric the conversion rate over performance yeah but the other metric that's just surged since the last time we've talked and it's so important is revenue per visit okay that is up 37 because they were converting at a higher rate but it was a lower lower click-through sort of thing but it was a higher rate last year the conversion rate was underperforming non-ai source traffic and the revenue per visit was also you know lower it wasn't we weren't just seeing as much sort of that that volume right and from a revenue perspective but here that's sort of scaling from on a per visit basis so it has a lot of implications for how you can think about if you drive more of this type of traffic if you make your site more machine readable then you'll be able to you know see a higher revenue per visit which will net out in higher average order values and overall revenue.

49:25And to help me understand a bit more also about the, is all this happening on the, if you like, the column out in the wild sort of gen AI answer engines, as it were, or is it this also that in combination with brands own sort of agents as well? And when that marriage happens, then the navigation is much easier. So this is purely, at the moment, third-party platforms. Okay. Plus machine-readable kind of like catalogs. Exactly. Exactly. And then they're finding what they need, and then they're making their way over to the site. So there is a stronger upside if you can also create a concierge-type experience, be able to provide answer engines on the site and experiences like that.

50:17But what we're even able to tap into is just this natural tailwind that many of these brands and companies are going to be able to enjoy from traffic coming from these types of sources because it's done a lot of the back and forth work and it's really helped qualify your audience before they've even hit your site. So there's a lot to unpack there and it's pretty wild to see these metrics surging this way and we talked a year ago this is only a two-year span that we've seen this change so if we've seen this much in two years you know over the next five to six decades it's only gonna be you know more impactful over time yeah i mean i've always thought and i remember doing this this this i wrote this kind of piece which was a bit of a uh i don't know a bit of a rant if you like call it an expose if you want to can use kind of like fancy words and it was that i went across identified a whole bunch of top retail sites and went into their their search bar on each kind of site and typed in um customer service into the search bar the the results were ridiculous i think on one which was the most comical, I got the recommendations for about, I think it was something north of 18 different types of white sneakers or trainers.

51:41And just, it didn't quite kind of, and it felt like a real opportunity, kind of like kind of miss. I think that almost felt like enterprise search was always a missed opportunity, but it feels like the, the, the, the, the genitive AI search assistance as it were i start to help kind of that sort of like stuff and and and is as really helping consumers and and the numbers are are proving kind of that that it's working and to that point specifically we we looked at null search rates on on-site search okay and the null search rate is improving so you're starting to see less incidences of users searching and then getting zero results.

52:24So you're having, again, consumers more aware of what they're looking for, what they want, but then you also have this search becoming AI-powered search on site. So it's able to contextualize so much more. It's not just doing a very kind of exact match, kind of, you know, keyword-based logic. It's really understanding the full context and then providing the users with more meaningful results. Perfect. So I wanted also then, because I know that you were speaking earlier today and I watched the session and there was lots of kind of good stuff and lots of data and lots of research in there, but I wanted just to zoom out a little bit because you highlighted three big trends that you're starting to see.

53:10And I think that might be, they might be kind of informative for the people that are listening in. So could you describe the three big trends that you're seeing and kind of what they sort of mean in broad terms? So I think we see a couple of big trends that we're profiling very closely. So the one key one is we continue to see that explosive growth from a traffic perspective. And that's just indicative of consumers wanting to lean on LLMs and Gen AI for more day-to-day activities. So that's a trend that really is the baseline for everything else that builds on it. So the fact that we're still seeing triple digit growth, not just across retail, but in travel and financial services and technology, it's really clear that this has a comfortability factor.

54:03So you're going to have more users just trying it out. And the ones who do try it out are going to be satisfied with their result and they're going to use it more. So that's one key piece that's really important here. The other important trends that we're seeing is just consumers really shopping with intent and really informing how they spend. So that helps inform the metric surges that we're seeing. So again, the conversion rate over performance, the revenue per visit is pretty sizable. and we're also seeing them spend more time across the site and the bounce rate is better. So all of this helps brands and companies.

54:45I think from our view that there's also a question about readability and how readable is your content. So another trend is improving readability and we're seeing that kind of incrementally improve, but we do see a breakout of certain top performers or the brands that drive the most AI visit share overperforming in terms of readability. We're in a state right now where about a third of the content across retail pages is unreadable. And from a machine standpoint. So improving that machine readability will help these brands drive more visibility. So that's important in terms of strategy and optimization.

55:30and so those are kind of the key things we're seeing in the data and i think what we're also very cognizant of is and one of the things we talked about in the sessions where you have perceptions around ai that you know kind of vary and shift but when if the broader utilization increases we do see that it's it's delivering on the outputs that these users are expecting and so that's helping inform everything that we're seeing across the board so let me ask a question because there's um the thing that strikes me is that technology has always changed fast and it's but it's it's the the pace of change is sort of accelerating has accelerated over the last few years but consumer behavior sometimes doesn't change as fast they don't build new patterns of behavior new habits kind of that fast and the thing that i was thinking about um recently was the idea of that that what happens to shopping in inverted commas because well so i know that people are are moving away from sort of searching by keywords and actually now searching by actually by intent or by mission or whatever and then that's because you've got an interface that can try and interpret that um but what happens to this almost sometimes a national pastime or a national kind of like sport that is shopping people just like rising around yeah doing their thing what's your data telling you about kind of that are people still doing this are we seeing just two yeah two different activities yeah i mean i think you you have still audiences out there that are comfortable shopping you know in traditional ways and traditional channels uh we we do look at this from like you know age and income cohorts and all that we do see this generative ai utilization for all these use cases kind of over indexing with millennials and gen z but we don't see gen x and boomer generations that much further behind so you do get the sense that many across the age cohort spectrum want to kind of test this out and experiment it and not potentially get left behind by some of these things.

57:49So that's one part of it. But you also have the LLM operating as a true informational source. So if you are interested in like, I want to go to a market that has little trinkets and I can shop in person, you can even lean on the arm to give you guidance around that. So as long as the operationalization of this continues, you're just going to see some of these, you know, it's kind of like how we think about personalization, where there's going to be different tastes, different modes people want to operate in over time. And it just becomes important for these brands to meet the customer where they are and adapt quickly so that they're not, you know, flat-footed when consumer behaviors and trends shift.

58:38So we talked a lot about what the research is showing and also about what customers are doing and what are they leaning on these new tools and what's happening. Here's another big question because you talked about one third of websites are not machine readable as it were. Broad, big question. Are brands ready for this? And if not, how do they need to pull their socks up? So we do have, as I mentioned, top performers and underperformers. And so that very, you know, level of readability, you know, can span across different enterprises. And you do see some brands leaning into leveraging LLM optimizer tools, trying to understand their readability and looking to optimize.

59:21And then you have an opportunity there across the board for brands to think about overall visibility strategy. So when you think about readability, you have to really think about structured content, structured text. You have to think about how to make this text very clear and machine readable so that it's your... The LLM has an expectation from the user to provide the right result as quickly as possible, right? And make it as correct as possible. And so making sure that content is absorbed very quickly by the agents is critical. So you're in a scenario where you have maybe too much JavaScript running across your site, and maybe you have too many elements that are kind of blocking loading and things like that, that can rupture the experience a little bit as the agent is ingesting all this content, that's something you're gonna wanna optimize.

1:00:22You're gonna wanna make sure that you're leaning on every surface to provide rich information back to the agent. So FAQ pages, help pages, store locator pages, this is all fair game to optimize. And so that becomes incredibly important. And then the broader strategy around how do you shop as a brand in the in the the ecosystem you know are you do you have reviews that are current are are there review sites that are you know in consensus with one another and kind of validating things and are we also seeing that when you're showing up it's all current is it's what's your social conversation looking like about your brand so when the user is asking questions like you know what's hot what's what's the latest trend in genes that i should be you know it should be able to absorb and triangulate across all that information to give give the right guidance back to the user so it's it's a combination of again the technical tactics across your web surfaces and then also showing up as a brand in a really meaningful way so you can get picked up on multiple fronts so if i was to asked you to that was a long answer to you know and kind of a complex kind of answer but if i was asked you to boil it down yeah because x says do these top two things what would they be i think the key is machine readability and making sure your content is well optimized for agents to absorb it making sure there's really strong structured content there and you have a tool for that and and we do and the the other opportunity is to think about how you show up as a brand across multiple surfaces.

1:02:12So by doing that, plus using the LLMO plugins, the optimizer tools, that will put you on a really strong footing to getting that visibility. Awesome. Anything else I should ask you? I really enjoy our conversations. It lets me get into a lot of different areas. I think what we're seeing is also just how quickly things have shifted in that yes the ai piece is there but when we also think about discoverability we're seeing social media influencers really over convert in terms of conversion and you know the discovery piece is what all these brands need to really re-understand and it's it's going to take some time because you're exploring new muscles of how to think about engagement but in a world where your your customer base is dealing with a lot of distractions a lot of things vying for their eyeballs the fact that there's so much potency in some of these areas the ai um generative ai platforms and then also social discovery with with these influencers commanding these massive audiences and then providing them with additional discounts and things like that all that becomes very important because these are relatively new engagement vectors.

1:03:34And so that's where the brands really need to hone in. Awesome. Thank you. Yeah, that's great. Always a pleasure. Wow. I really enjoyed my chat with Vivek. He always brings the research and data. Love a bit of research and data. But I particularly liked how you described how Gen.A.I. traffic is highly qualified and now high value. It's a whole new world out there. But before I start moving on that, let's bring things to a close. So I really hope you enjoyed the series of conversations that I had whilst at Adobe Summit event last week in Las Vegas. And it gave you a flavor of some of the things that are being talked about and gave you lots to think about and you learned a lot too.

1:04:11But thanks again and do tune in again. Thank you.