AI agents - personalisation, productivity & performance with Adobe, ServiceNow and IBM

0:06Ladies and gentlemen, welcome back to a special edition of Uncensored CMO. Now, this is brought to you in partnership with Adobe. Everybody is talking about AI and how is AI going to impact on marketing? I was fortunate enough to go to Adobe Summit where AI was the main topic of conversation. The advances in AI just in the last six months are, quite frankly, mind-blowing. So while I was there, I got three guests to come on the podcast and tell me what AI will mean for the future of marketing. And they are three spectacular guests. We have Colin Fleming, the CMO of ServiceNow, Billy Seabrook, the Chief Design Officer at IBM, and Stacey Martinet, who runs communications for Adobe.

0:42And I put all the questions I've been wanting to ask about AI and how it impacts marketing to them. Fascinating episode. Three guests. You're going to love it. Here it is. So here we are, Uncensored CMO in Las Vegas, and I'm joined by none other than a returning guest, Colin Fleming. It's good to see you again, mate. How are you? This is amazing. We're in Vegas together. This will be a lot of fun. It is indeed. And you're doing a big keynote speech tomorrow. I'd love to get a bit of a preview if you can. so happy to you yeah what's what's it going to be about well it's an incredible forum obviously um you know marketing to marketers is one of my favorite things um so we're going to take the opportunity to call a little bit of bs on marketing a little bit i believe that b2b marketing we've talked about this in the past we've kind of confused ourselves a little bit right the long and the short of it we've talked about peter and les and the folks like this but we have to kind of reset a little bit on b2b marketing so it's just going to share some of the lessons that I've learned along the way.

1:36So five lessons around how we're really thinking about playing the long game, thinking about investing in buying groups, thinking about talking about, you know, true personalization and really just changing the perspective away from leads and into a bigger story. Yeah. And really speaking the language of the customer. I think that's important. You've obviously got a lot of experience for B2B marketing. What are the things that people don't understand about how marketing works that they should do that would give them advantage? Because if i just think about my kind of system one experience like most b2c marketers well versed in you know long-term short-term role of emotion all those kind of things b2b marketing seems to be a little bit more kind of service-led rather than brand pull if you know what i mean so yeah yeah what are you finding you know i think b2b marketing and i i'm complicit in this so i'm a b2b marketer as a career so it's all good but i think a couple things are true number one is we market attribution models far too much.

2:28Like, you know, the average deal cycle at my company is over 400 days with a 90-day attribution window. You're not getting the full picture. And so really, really separating the way humans buy products from the way you measure it, I think too often they're one and the same, but really being able to separate those two things to a degree is important. So number one there, I think, you know, we market with our org charts too often and the systems we use put a lot of friction in the system. that doesn't really really work um and i think it's about removing friction we actually spend a lot of our time marketing for credit instead of marketing for actually impacting us yes which is no what nobody wants that and then one of the things that i want to talk about is how b2b you know none of the b's in b2b stand for boring we can be opinionated we can have we can be culture shaping we can have a point of view we can be creative and you know 70 or so of the gdp in the US here is generated by B2B businesses.

3:21Yet for some crazy reason, we've decided that our bar can be lower, held to a lower standard. Does this make sense to me? So frankly, I kind of want to challenge it tomorrow at the keynote and see what happens. Yeah. Now there's a whole raft of announcements from Adobe about how they're investing in, you know, agentic AI is a new phrase that I learned yesterday. But the world has changed a lot, even in six to nine months, isn't it is change what are the big shifts that you're seeing in terms of the how ai is being used to help business well it's approachable now obviously it started out as a we were ai curious to a degree you know you're playing with firefly and you're placing the incredible technologies that adobe has we moved it to be more of a mission critical perspective so we're implementing things like segmentation and persona development and category entry points with synthetic research We're doing AI agent orchestration from ServiceNow agents to AI agents from Adobe to really get the best of both worlds.

4:16Like those types of scenarios we weren't even talking about four months ago. And it's all real now. So I do think that the ability to scale brand and govern brand has never been as accessible as it is today. And really allowing these agents to handle the sort of mundane, if you will, so we can focus on the strategic lens and the personalization, the creativity, and the curiosity. that's what it really unlocks and that's where we're going to spend our time. Yeah, yeah. And I think everyone fears, don't they, that the role of the AI is to replace jobs. But I think what you've done at ServiceNow really helpfully is to add people back into your position and go that they're working for you, which I think is a really good...

4:51Put AI to work for people. Yeah, exactly. And you've got a new campaign out that's been running for the last few months with Idris Elba as well. He's a blast. We're having a lot of fun. But again, it's trying to take this crazy technology and not make it about the technology. Yeah. And this is what B2B marketers, I think, do, back to your original question, which is you see a lot of AI agents being marketed as the bleeding edge technology. But I think what we're trying to do is find a way to soften it, to make it more human with put AI to work for people, calling the agents personal minions, put it in scenarios that are more approachable.

5:22And I do think that if we can come across as the human side of AI with some killer technology, I think we'll be in really good shape. Now talking about Adobe, of course, they've got a suite of AI products. You have as well. and we talk about orchestration as well how do you integrate all that technology together what's your relationship with Adobe yeah well first I'm a huge fan of Adobe they've been in my 10 months in this organization they've been my ride or die trusted partner and they're so much more than a vendor to me they're absolutely advising us on how to implement buying groups how we should think about the future how we should think about AI agents and in turn we're actually a design partner for a lot of their new technologies whether it be buying groups and those AI agents as well so there's a very much a back and forth technology sharing if you will we like to play with their bleeding edge technology and give them feedback on it so that's a lot of fun and in turn we're just getting the right advice and from people that you know this is an organization that bleeds marketing this is what they do and so why not learn from the best yeah makes a lot of sense so anyone listening that's kind of like you know slightly that days by you know everything that's going on what would be your advice to marketers to kind of intend to you know catching up with what's going on in AI and making sure that they're using it in the right way, not falling into some of the traps they might fall into.

6:35Yeah. I mean, invoke the curiosity. Find every scenario you're in, whether it be data, whatever side of the brain you're using. Ask yourself the question, how can I be using AI? How can I be using AI? And finding incredible ways of unlocking this. I was commuting down to Santa Clara the other day and I put chat UBT in voice mode and I talked to it on the entire ride down, just asking questions about how should I be thinking about this? where is this heading? How shall we think about positioning? How shall we think about segmentation? You name it. And just finding the curiosity that we have to unlock.

7:07I think that's the big, big, big thing. And of course, don't be scared of it. It doesn't bite. It's not always perfect. There's hallucinations. There's those things that exist. But those that are curious and those that implement it and those that experiment with it, they're going to be the ones that come out ahead on this. And so I'd rather be on that end than the one that's forgetting it. It's interesting you use it for voice, actually, because it's a different dynamic. Because when it's in written form, You're going, oh, that's not quite right, or that word seems a bit odd. But in a voice, it's like conversation where you're exploring something with someone and going, oh, I didn't mean that.

7:36I meant this. That's right. You know, it's a different experience, isn't it? If you haven't done this yet, it's kind of mind-blowing, actually. I encourage you to do it maybe on your flight. Yeah. I look like a weird person on the subway sometimes talking to myself, but it is fascinating. And again, I'm challenging myself every single day just finding ways to be more curious. Yeah. And I do think that's an unlock for people here is finding the curiosity with Adobe technologies, with ServiceNow technologies. What other scenarios should we be thinking about? Because the technology is there. We have some work to do on data.

8:05We're all suffering with a little bit of lack of data quality from this. But the opportunity and the technology is there. We are truly only limited now by our curiosity. The data is a very good point, though, isn't it? Because you need to be training the tools on the right data set to get the good answer out. It's the fuel. And if you put poor fuel in anything, you're going to get poor out, right? So poor racing analogy, but there it is. Brilliant. Well, listen, looking forward to your keynote tomorrow. It's going to be very good. And yeah, I look forward to the reaction it gets. Awesome. Thanks, man.

8:32Great. All right. That was Colin telling us about how they are using AI ServiceNow. And now we go over to Billy Seabrook, who is the chief designer of IBM, to find out a bit more about how IBM are using AI and how they're partnering with other companies to implement it as well. Here we go. So, Billy, welcome to Uncensored CMO. It's great to have you here. Thank you. Nice to be here. um just uh if anyone listening watching just introduce your yourself and what does the chief design officer do sure yeah well i'm billy seabrook i'm the global chief design officer for ibm consulting been with the company just under eight years um and i have a background in marketing digital product uh engineering um change management you know and brand building uh and and design of course um so worked at companies like a bunch of agencies ebay Citibank and others.

9:18Now, everyone's been taught about AI for the last few years, and I'm really intrigued to get your perspective on kind of where are we on the journey. Now, in my day job at System, when I get to work with tons of technology companies, and it's really fascinating that everyone has got an ambition to be leaders in AI because they can see where the kind of ball's going next. But actually, most of them are on this kind of discovery phase where they haven't quite worked it out yet. And I'm just intrigued to know, I mean, given the role you do, where do you think we are on the journey? That's right. Well, I mean, it is moving so fast.

9:45And I think some people are early adopters and some are a little bit laggard and some are in between. We've done a lot of research on the topic. We've released a few studies from IBM, actually a bunch of studies from IBM through our research arm, the Institute of Business Value. And we actually just launched one here at Summit this week. It's called the Content Supply Chain AI Awakening. And it really does probe on people's adoption of AI in the content creation process and sort of where they are in terms of what their aspirations were. And what we discovered is that for all the people that have answered the survey, and we're talking 1 ,100, 1 ,200 people, that only 50 % of them have actually reached the goals that they had anticipated for adopting it.

10:28However, over 80 % think that there's a huge advantage to adopting it. So they know they want to do it. They know why they should do it. They know what to do. It's a little bit of not knowing how to do it properly and responsibly is where it gets hung up. And actually, the trend is growing. So over 60%, almost two-thirds of the people think it's only improving, that there's even more endorsement of AI in content supply chain. It's just getting there. And the risks behind it are kind of the usual suspects you hear. So it's trusting the AI output, the risk around legal and compliance. Actually, change management of the workforce is a big one that spiked.

11:08And I think that's maybe underestimated in terms of just retraining your people into new behaviors and helping them along that journey, especially if they've got to change their entire skill set. And there's cost still. Even though the cost of AI is going down rapidly and people are adopting more small language models instead of large language models, there still is cost there for marketers. Yeah. Now, in terms of the use case for using AI in creative development, what are the principles of a really, really good and effective kind of campaign output? Well, I think authenticity is key, you know, and accuracy.

11:44So there are good use cases for AI and poor use cases for AI currently. Anything that requires this level of accuracy that's going to actually have a legal compliance to it, or you're, say, launching a product that you can't just take risks on the generation of that image, you're going to want to do more classic image generation for that type of work. But if you've got the flexibility to change out more abstract backgrounds, fantastical type of imagery, it's actually getting better and better in terms of realistic lifestyle imagery. So if you want to extend, use it for versioning and resizing purposes, like the end production phases of a project, you're going to start to get huge productivity gains, and you can trust the output.

12:27But I think authenticity is key, And that's where you sort of get the spectrum of possible use cases. But I also think personalization is huge. You know, that's what everybody gets excited about, personalization at scale, because of the volume of content you can now produce instantly. And that means translation. That means obviously resizing for different media formats, changing out the backgrounds for the context of maybe the geography you're talking to or maybe even the persona you're trying to reach. Those are those great opportunities that excite people because we know for a fact that more personalized content that is contextually relevant for the user is going to drive more engagement.

13:03Yeah. Now give me some data on this because you've been running this for a couple of years now. Yes. On some of the IBM campaigns. What's the data telling you? Yeah. So we were quite early adopters at IBM, especially in our marketing organization. Of course, on my side, I work with clients on their adoption of AI and their campaigns. We actually have been on a journey for over four years with Adobe. It's important just to set the context of how we've built trust in the technology over the years. But about four years ago, we realized we needed to simplify our marketing stack, simplify our marketing period to get more ROI, be more cost-effective, drive better results.

13:36And we went on a four-phase journey. And the first one was just getting our data in order, you know, putting it, you know, centralizing the data, deduping, things like that, cleaning it up. The second big phase was moving, simplifying our stack and going to the Adobe solution, which took out huge amounts of cost right then and there and just made it a much more efficient CMS and use of assets. Then we went to phase three, which was really the change management, reorganizing the entire marketing organization, being much more simple, allowing more nimbleness, and again, building confidence in the tools, which then allowed us to jump to phase four, which was experimenting with AI.

14:15So when this all sort of, and of course, for context, IBM is an AI company. You know, we are in the space. So we've got probably a little bit more experience and faith in the technology. And Adobe uses your AI as well. That's right. And Adobe uses our AI in their technology as well. So the underpinnings of AEM is actually there's some Watson AI in it, as well as the Red Hat OpenShift technology that allows them to put AEM on different clouds, which was just announced today. So about two years ago, rewind, two years ago, Firefly was just announced. It was effectively beta. You couldn't use it for commercial uses.

14:53And it was an asset here at Summit. And our CMO and the marketing team were blown away by the potential of using AI to generate hundreds of assets that would take a designer and a creative team a very long time to do. So we had the Masters Golf Tournament coming up. And we wanted to do a big social media campaign. So we got special permission to use Firefly to generate the campaign. And it was a very simple concept. It was just a question mark. It was the What If campaign. It was supposed to be provocative to get people thinking about What If. And we just did hundreds and hundreds of versions of this question mark textured in golf-related themes.

15:32We released it on social media, and it was the best-performing campaign in the history of IBM. It had 26 times engagement, click-throughs, sharing. And you can probably attribute it to the fact that there's such variety and such personalized and contextually relevant creative that we put out. So that gave us a huge boost of confidence. And we then since experimented more and more in our content creation. We've built AI tools ourselves to accelerate the whole content supply chain for IBM. We're very excited about the Gen Studio announcement because, I mean, they're really starting to connect the dots in terms of the two clouds of Adobe.

16:12But yeah, last year we did the next wave. So we understood that generative AI is great for the production, the long-tail production. That's 80 % of the creative process. Huge cost savings, huge productivity savings. That's what we did with the masters. Last year, we actually tried to go the other side, the ideation phase of the creative process to see if generative AI could give us new inspiring ideas and accelerate that process. And we launched the Trust What You Create campaign here at Summit last year. commonly referred to as the fishy fish, fishy AI campaign. And we used generative AI to come up with, if you're unfamiliar with the campaign, it's effectively a bunch of fantastical fish swimming around in a fishbowl that represent AI that's gone wrong.

17:06You know, so untrustworthy AI. And the idea is that Watson can come in and clean up all the the fictitional or the mutant fish and turn them all into beautiful goldfish but to come up with those mutant fish concepts of you know and they're all various characters and it was very conceptual and funny we used firefly to generate those characters um so incredibly it was an ideation booster creativity booster and it accelerated our workflow to actually give those models to a real cg engineer and actually develop the develop this now we're having this conversation in Vegas, aren't we? And they appeared on The Sphere, didn't they, as well?

17:46Exactly. So that was a big media placement, for sure. We took over The Sphere. It was perfect. And I said The Sphere, we knew we were going to do that, so that inspired the creative concept, obviously, of turning it into a fishbowl. But it was a highly referenced campaign. People loved it. And again, huge, huge business benefits for us. Not only did it just get great word of mouth and it communicated our concept, but it saved so much time. It just wouldn't have been possible normally to do a campaign of that magnitude. And all the versioning that happened. Well, listening this morning to the keynotes, they were talking about, on average, a 10x improvement in speed.

18:24That's right. And a 30 % increase in click-through rates, which is astonishing. It is. But presumably that has pretty profound impacts for your organization as well. What do you see as the next step once this becomes the norm? I think there are two things that are happening. One is this incredible productivity boost. Just assuming you've got all the governance in place and you still need quality checks for all these, a lot of the more, you know, hero assets and things, even the derivative assets, you really do need trained professionals doing quality checks. But still, it is like 10x of improvement in productivity, which means what do you do with that banked time?

18:59You know, do you just move on to the next project? Do you work with less people? Or do you actually open up the possibilities? you can do two things. You can actually take that time and invest it in some of the early concepting and research phases of a project that usually gets short shrift in a project. It's all about coming up with the idea as fast as possible, spending the time to produce the hero asset, and then all the production on the tail end. If you can eliminate or at least reduce all that production, you can look at all the thoughtfulness that can go up front. So that's a good distribution of time.

19:33But you can also, instead of just doing one campaign, now you've got all this headroom, you can do five campaigns. And you can get more thoughtful about the personalization and the localization of those campaigns so they're more effective. So that's sort of one thing of like, what do you do with all this time? The other area is sort of skill set shifts too. And that means both the role of a designer, being less on the production side, a little bit more on the conceptual direction side is a big shift. And also including more people in the creative process. So democratizing the whole creative content supply chain, allowing more field marketers to do their own versioning of work with tools like Express.

20:12But it's like any change management. People have to go and learn new skills, change their skills, get used to new kinds of ways of working. And the third thing I would just say is we did a study actually last year on the impact of AI on designers specifically and marketers. And we realized that there were some blind spots from some of the respondents saying, like, what are the most important skills that you need? Hard and soft skills in the age of AI. And creativity boomed at the top. Speed of delivery was huge. Decision making was huge. These are like the skills that people really value now. But on the very tail end that people weren't thinking about were empathy and caring about the end users and the impact of what you're doing.

20:55and that was a bear that nobody thought that was important and i think that's a huge blind spot we're going through a massive change people have to be open and willing to learn new things but also not um not forget that good you know professional craft skills are still really important yeah and that i think is where the debate is do you see it as a threat and is a cfo just going to use it as a cost-cutting exercise or is an investment where actually you get more time back to be creative more times about to be spontaneous that's right be reactive and there's a huge opportunity for people that want to take it that's right yeah and that's and that's comes down to company culture yeah your your you know your business goals you know every company is different and how they you know square that equation of where do they where do they adjust you know and take and unlock new forms of value yeah i think at ibm specifically we are definitely redistributing some of those savings we're definitely banking some you know cost savings were great we're definitely moving faster which is great but we are redistributing some of the time to be more thoughtful in places that never we were able we didn't have the time to do it and quantity and personalization of the campaigns is huge for us yeah amazing billy thanks for thanks for sharing your your time and expertise on this it's a fascinating subject and it's great to talk about uh how you're using it at ibm all right last but not least over to stacy martinet now she has been running adobe and their communications and put the event on she knows an awful lot about ai so i wanted to catch up with her and find out what it means and what adobe you're doing, particularly as they are marketing to marketers and they are using their own products themselves as well.

22:26Now, I should just say quickly, this was recorded in the conference itself. So there's a bit of background noise. So hopefully you'll forgive that, but the content will be worth it. Here we go. So Stacey, here we are at Adobe Summit and this is part of your role, isn't it? But tell me a bit about your role, how long you've been at Adobe and what have you been up to? Yes. Well, thank you so much for coming and thanks for having me here. I really admire your podcast thank you you've had a lot of people i respect and also a lot of friends uh so it's an honor to be here adobe summit is adobe's annual digital experience conference we're here in las vegas and part of my role is i lead events communication social media so that all rolls into one to come to life here here at summit now the slightly funny thing about your role is like you're effectively a marketer and you're marketing to marketers yeah which i just think is is quite quite a wild idea what makes great marketing for marketers it's very meta yeah it is so you know my first job was i worked in communications and public relations at a newspaper so you had to do public relations for journalists yes which was also very meta and maybe the things that they have in common is that what makes great marketing for marketers is like the story and simplifying the power of what Adobe brings to our customers.

23:47Yeah. We spend a lot of time in the marketing industry talking about tools and the craft to deliver to end customers, right? Yeah. Consumers, or if you're B2B businesses, but we don't talk a lot about the marketer, right? The person who has to go into the boardroom every day and defend the effectiveness, the efficiency, do more with less. And so I love that that's what we do at Adobe And that's what this conference is about. Now, I love the fact that you said simplicity, because actually the thing that's like I find completely incredible is the pace of change. Yes. I mean, I've been learning loads of new words this week.

24:23And it's, you know, AI particularly is in a completely different place to where it was six months ago. So it must be like a big challenge that how do you kind of, you know, communicate all this change to marketers and help make sure that they're kind of up to date? Well, it's all about AI. Absolutely. And you're right in that when you think about even three months ago or three weeks ago, what was the news driving AI and what was going to happen next? It is very fast. And part of it is the models, right? These data models that power AI are rapidly evolving. And so the architects and the engineers and scientists are finding more and more use cases that they can be applied to much more quickly than they were two years ago.

25:09Right. Because that's the usage that informs the model. So you get this cycle that's like the more people use it, the quicker you can actually innovate with it. Now, one of the big things for me was the difference between, you know, AI being deterministic, as I think is the right phrase, isn't it? Correct. And then agentic AI and understanding that was a big revolution, isn't it? And how that turns out. Can you just explain what that is? Yes. And the use cases that you described. I talk about original AI. Yeah. Which, you know, we've been, AI has been around for a long time. Adobe's been working on AI and it's been in our tools like neurofilters and Photoshop or liquid mode and Acrobat where you can really zoom in on a text.

25:47That's AI and that's deterministic. But the breakthrough was around reasoning. So agents can reason. They can take the data and they can start to make decisions. And agents, how we define agents, are AI that can work with you. They can communicate with you as a human. But they can also communicate with other agents. Wow. Which unlocks a lot when you think about what's possible in your workflow and making you work smarter faster. So what we announced was a series of agents through our Adobe Experience platform that you can use to speed up your productivity. Yeah. And which bits of the kind of marketing mix do you think this will impact?

26:26I mean, can it do everything or are there certain parts of the marketing kind of role that it will impact the most? Where do you think it's going to be felt? We're focused on the tasks or the areas of work that tend to take a lot of time, but they aren't inherently the most strategic or the most creative. Right. It's not where you're doing your best work or where you want your talent focused. right and so when you think about things like audience segmentation you can now use a certain agent to go in and say i have this product i'm trying to sell it at a certain price point based on past campaigns serve up these audiences yes right they can do it like that then you could say great i have hundreds of assets that i've run over the past few months can you go through and pull the content you think would perform best for these audiences.

27:16Right. And how long would that take without this agent? And you would be in multiple tools and maybe you'd be on your desktop, you'd be looking at your phone and now the agent can do it. So the agent works on your behalf. That's one part of it for the practitioner. What I think is really exciting. And we talked about with brand concierge is for brands to be able to develop agents that they interact with their customers. Oh, interesting. I hadn't picked that bit up. That's quite interesting. So you can sort Or delegate the customer service responsibilities, solve their problems. Yeah, and think about it like the next level of customer service and experience, where, say, you're a Marriott Von Boy member.

27:51There's a lot of things Marriott knows about you, right? The cities you've been to, the restaurants you've eaten, the classes you've taken. And so now it says, oh, you're coming to Las Vegas, John. Do you want to stay at this hotel again? Right. Last time you went out and you had Mediterranean. Here's a great restaurant that we recommend. And so it can start to truly personalize your experience. I love that because like we've been talking about personalization at scale for so long now, haven't we? It's been one of those kind of buzzwords and it feels like it's finally arrived. It does feel like the promise of personalization is upon us now.