Stay up to date on the digital shelf.



    We'll keep you up to date!


    Real-World Examples of How Top Brands are Using AI on the Digital Shelf

    AI’s impact on e-commerce continues to grow with many brands using test-and-learn approaches with AI in their content creation, testing, syndication, and digital shelf optimization practices. Those who aren’t taking advantage of AI-based tactics risk falling behind the competition, but they’re surrounded by questions about how to get started with AI and what the new era of content creation will look like. In this audio rebroadcast of a recent Digital Shelf Institute webinar hosted by Lauren Livak Gilbert, learn how fellow digital leaders like Todd Hassenfelt from Colgate and Pam Perino from Ghirardelli approached the same questions. Eli Orkin, VP of Marketing at Vizit, also joins to map out how AI-fueled data can drive measurable performance improvements on the digital shelf.

    Peter Crosby (00:00:00):
    Welcome to Unpacking the Digital Shelf where we explore brand manufacturing in the digital age.
    Peter Crosby (00:00:16):
    Hey everyone. Peter Crosby here from the Digital Shelf Institute. AI's Impact on E-Commerce continues to grow with many brands using test and learn approaches with AI in their content creation, testing syndication and digital shelf optimization practices. Those who aren't taking advantage of AI-based tactics, risk falling behind the competition, but they're surrounded by questions about how to get started with AI and what the new era of content creation will look like. In this audio rebroadcast of a recent Digital Shelf Institute webinar hosted by Lauren Leach Gilbert, learn how fellow digital leaders like Todd Hassenfelt from Colgate and Pam Perino from Ghirardelli approached the same questions. Eli Orkin, VP of marketing at Visit also joins to map out how AI fueled data can drive measurable performance improvements on the digital shelf.
    Lauren Livak Gilbert (00:01:06):
    All right, thank you everyone so much for joining. We have a big audience today. I think AI is a very popular topic for everyone. I'm really excited to be here with all of my guest speakers with Eli, Pam, and Todd. So welcome to real world examples of how top brands are using AI on the digital shelf. I know this is a very popular topic and we are going to go through some specific examples and even have a bit of a fireside chat. So for those of you who do not know me, my name is Lauren Vac, I lead the Digital Shelf Institute. I'm really excited to see a lot of familiar members and names on the attendee list, and I'm so excited to be welcomed today by so many exciting guests. We have Eli from Visit Todd from Colgate, Pam from Ghirardelli, and I'm going to let them give just a quick intro before we get started. So Eli, do you want to kick us off?
    Eli Orkin (00:01:57):
    Sure. Thanks Lauren. Again, my name is Eli Orkin. I'm a founding team member here at Visit. And just quickly, for those of you who are unfamiliar with Visit, we're an AI company that builds software to really empower the measurement, monitoring and optimization of visual content to drive conversion and performance both online and off.
    Lauren Livak Gilbert (00:02:19):
    Awesome. Thanks Eli. Todd,
    Todd Hassenfelt (00:02:22):
    Hi everyone. Thanks for joining. I'm Todd Hassenfelt. I'm the Global Digital Commerce senior director of Strategy execution at Colgate Palm Olive. I'm lucky enough to work with a lot of talented people. I'm on the global digital organization and Colgate is a caring, innovative growth company and really excited for this conversation.
    Lauren Livak Gilbert (00:02:39):
    Thanks, Todd. Pam.
    Pam Perino (00:02:42):
    Hi everybody. My name is Pam Perino and I work for Ghirardelli Chocolate Company. I'm the digital content operations and development manager in charge of content organization and also our syndication as well with our retailers. Looking forward to
    Lauren Livak Gilbert (00:02:58):
    Participating today. Awesome. And I love that it's complimentary brands, right? Because you're eating chocolate and then you should brush your teeth afterwards, right? So it's just a perfect combo. So thank you both so much for being on the webinar. We're really excited to hear from you. So I mentioned this a bit already, but we're going to jump into just a bit around how AI is changing the landscape. Eli will jump off on that. Then Todd and Pam are going to share a little bit more about their AI journeys and then we're going to have a fireside chat with both Pam and Todd where we'll ask a couple of questions around how they've been using AI and how they've been able to work through that inside their organization. But as always, on any webinar, please feel free to submit a question in the q and a or in the chat and we will get to it either during the presentation or at the end.
    And in traditional DSI fashion, we always like to have a poll so that we can get a sense of who we have in the audience. So if we could pull up that poll right now, Molly, that would be great. So if everybody could go back to their screens and make sure they answer this question. Are you currently using AI in your digital strategy? Yes. No, I don't know. Or we are in discussion. We have a lot of participants, so lots of results are coming in really fast and it will be really interesting to see where we land. So I am going to close the poll in 5, 4, 3, 2, 1. Alright, let's close that poll and see where we're at. So, okay, 35% of people said yes, they're using ai, 21% said no, 37% said that they're in the discussion and 7% said they don't know. So I feel like this makes a lot of sense. A lot of people are discussing it and trying to figure it out and some people do have it in their strategy and some don't. So I don't know from Pam, Todd, Eli, any reactions to that or thoughts?
    Todd Hassenfelt (00:04:59):
    Oh sorry, go ahead Eli. Todd, go for it. Go
    Eli Orkin (00:05:00):
    Ahead Eli. I was going to say I think that's great that from these responses that so many folks are either already using AI or considering using it. I think it speaks to the innovative nature of this community and I'm sure we'll have some really awesome q and A at the end of this. And for those who are not, a lot of folks are still trying to figure out what tangible, what real business problems AI can solve and how people are implementing it. So this is definitely the right discussion for those folks as well.
    Todd Hassenfelt (00:05:34):
    Yeah, I was just going to say considering chat GPT kind exploded just about a year ago coming up here at end of November. I mean a positive side of it is look how many people are discussing or already are using AI in their digital strategy. And you would expect it takes a long time of adoption of a lot of things. So this should be a fun discussion no matter where you are in your journey.
    Lauren Livak Gilbert (00:05:55):
    Awesome, I love it. Alright, Eli, I will kick it to you to start us off.
    Eli Orkin (00:06:00):
    Thank you, Lauren. I think a major reason that we've started to see brands begin to invest in AI this year is that when it comes to digital content, especially visual content, traditional methods and mechanisms for analyzing that content, gaining insight into maybe competitive content and actually getting recommendations around how to potentially improve and optimize your content to drive that increased performance and conversion don't really scale when it comes to the digital shelf. And that's because these kind of traditional or maybe more manual methods of gathering that information, surveys, focus groups, panels, even AB testing, really aren't equipping teams with the intelligence that they need to keep up with the speed of today's digital consumer. And that's it's really because when it comes to those methods, they've got targeting limitations. You can imagine trying to hone in on your exact specific consumer audience or maybe test against all of your global brand audiences.
    The reality of modern buying behavior is system one, that quick, effortless emotional decision-making, not something that's really easy to replicate in a focus group, a clinical or a survey setting. They can be time consuming, it can take days, weeks, months to run some of these methods and then it can be really expensive to run and iterate and cost thousands of dollars to test just a really small set of content or visuals with a single consumer audience. So kind of makes it impractical if not impossible when you think about potentially thousands of products or thousands of SKUs over hundreds of retailers. So this gap in what traditional or maybe these more manual methods can offer and what's actually required when we think about analyzing and optimizing the hundreds or thousands of assets a brand might need to in order to optimize their content across digital. Again, nevermind those competitive insights too is why we're seeing more and more companies turn to ai.
    And that's really what I'm, and I know others on the presentation today are excited about talking about in this space is these advancements in technology and how AI is actually beginning to solve a number of these real business problems. We know that AI is already impacting so many areas of the user experience. If we think about our daily lives streaming platforms like Netflix and Spotify are giving us curated content based on advanced learning and predictions about what we're most likely to want to watch or listen to next. I know my Netflix keeps trying to feed me Love is blind, I don't know what that says about me, but on the digital shelf we're seeing take advantage of this when it comes to increasing personalization around product recommendations and offerings and even really the timing of those offerings. If you've had customer support interactions on a large online retailer recently, the odds are you've already interacted with AI as it's fueling customer support chats and experiences, especially with the amazing progress. Todd mentioned this, but with chat GPT and other large language models recently. And then even when it comes to forecasting sales or making decisions around inventory, we're seeing new tech emerge that helps users make more informed quantity orders, understand the right products to potentially prioritize and even save time when it comes to that purchasing or repurchasing.
    And then recently we've seen how AI is even changing the way that we edit content with generative fills tools like Adobe's Firefly, even the way that we create new content with prompted image generation on platforms like Midjourney and Dali. These programs can help us increase efficiency for sure when it comes to making some routine edits or maybe removing elements from visual content, adding in elements or even provide us with completely generated visual content when we might not have access to photo shoots or stock imagery. Just speaking for Visit, our visual AI is unique in the sense that we're empowering our users to really see content through the eyes of specific consumer audiences. So we do that by harnessing a really large scale analysis of organic interactions with online imagery that consumers have every single day to teach our ais to measure imagery for appealing and unappealing characteristics of content and really what aspects of that content will best capture attention and drive conversion. So the outcome here is basically an AI powered simulation of those consumers visual preferences that can be used to test content in real time gain insights around how to better present product content to shoppers measure the health of your content across digital. I don't want to steal any of Pam or Todd's Thunder here, so I'll leave a lot of that to them, but they really have some innovative and powerful real examples of this application of the technology that they'll be covering a little bit later on.
    And then just finally, just to add a little bit more color here, this is a recent survey that was commissioned alongside the Path to Purchase Institute regarding AI prioritization around digital content in 2024. You can see when it comes to what e-commerce content and digital content professionals are currently most likely to already be using AI for right now, that's overwhelmingly that generative AI images, video music. But in the next year when it comes to what professionals and brands hope or intend to leverage AI for, we're seeing a shift from that generative to ongoing performance measurement, advanced analytics to optimize content performance and content personalization and customization. So from what we've seen, this is partially driven by that learning curve around ai, what it can actually do, the problems it can solve. For example, generative AI is amazing. The ability to spin up hundreds of visual content options at the click of a button is huge game changing. But now we're starting to see folks ask some of those next natural questions in that progression, is this content the right content? Is it effective? Will it capture the attention of my consumer audience? Will it drive performance? And if not, what can I do to improve it and measure that content appeal with my audience over time?
    And then finally, this is probably not a huge surprise given the kind of mounting interest in ai, but teams will have the support of their C-Suites. Respondents to this survey also indicated they expect to see their budget allocation for AI-based tools triple over the next year. So the question is how do we begin here? What are some of those major challenges you may face and how do you solve real problems through applications of new AI tech on the digital shelf? I'm going to now pass it to someone smarter than me, Pam Perino who runs again, digital content operations at Ghirardelli. She's an incredible thought leader and an innovator in this space, so I'll pass it over to her here.
    Pam Perino (00:13:39):
    Well thank you Eli. That's a wonderful introduction. Hi everybody again, just real quick, I'm Pam Perino and I work for Ghirardelli Chocolate Company. I've been with the company a little over three and a half years and absolutely love it. I love the company, I love what I do and I love the CPG space. So our focus area really is especially this year, really looking at how do we innovate and create and be as quick and nimble as we possibly can with current tools as well as looking at opportunities with new tools. And one of the first ways we integrated with AI is with visit in looking and measuring our images, really looking at how to measure our current image carousels, our hero images and looking at opportunities to make them more effective. We're a small but mighty team here and so having some of these tools really, really helps and create better ways to work not only for myself but my cross-functional teams as well.
    Cross-functional collaboration is absolutely critical. And before we really started this whole journey, it's really a matter of getting to all of your cross-functional teams and partners and talking to them about ai. I think for some teams it's a little bit more intimidating than others and some people it's like the wild, wild west out there, but I think it's a matter of harnessing these amazing tools and looking at what it can do for us and how we can continue to improve our content to be more relevant and engaging to our customers. One of the biggest hurdles we overcame is really probably the legal and regulatory guardrails. I've worked very closely with our regulatory team and all of my brand marketing team and everybody involved there, but being in the food space, obviously there's a lot of regulatory guardrails that we have to keep in place. And so one of the things that we talked about was really these tools and how do we monitor them?
    How do we make sure that we're following the same checks and balances as we always have and we'll continue to do so even using some of these new tools. So that's probably been one of the biggest challenges. I think one of the things that, again, I mentioned earlier strategy behind our AI image selection optimization. One of the big projects that we worked on a little bit earlier this year was we have some new baking items that we've been launching and we had some beautiful new photography from a photo shoot and I worked very closely with our art director and his team to really look at the overall images that we shot for each of the three items and then come up with several different iterations of crops so we could measure those in visit using specific audience lenses and are picking the right benchmarks and things like that.
    And then really looking at how to pick the right image that would really hopefully generate customer to not only click into the product page but actually buy and put the item into their cart. So one of the things we did discover, and it seemed kind of a general overarching discovery, was that the items that we kind of focused in a little bit tighter on a little bit of background and you could see the package on the left side, that one scored the highest, those seem to rank higher and seem to resonate better with visits AI tool, which is exciting to see and really helps us as we're looking at images and even planning for photo shoots that are coming up in the future. What are the needs for content for our e-commerce sites and retailers as well? So that's one of the things. And as far as our future AI plans, we've also just recently started working with citation and their very cool tool, rough Draft Pro to begin copywriting, which we're super excited about.
    So we're now working with AI also, not only with visit and measuring our images, but also in copywriting, which is great. We look forward to using the tool to again accelerate and help us optimize our product marketing copy for all of our key retailers and also to really making sure that we can quickly turn on and off seasonal content implementing say Halloween titles into our product in bullets and also in product descriptions. And then also too making sure that within our image carousels that we're being able to create images that are within our carousel that are very seasonally oriented, maybe flipping out one or two and testing those and visit first and maybe putting something in that's related to that particular season and carrying that through and then being able to go back to very quickly using a evergreen image. So definitely being able to move faster and quicker with these new tools is very exciting for us and got great support from our company, not only Ghirardelli but from our global group as well and these great tools. So we're very excited.
    Lauren Livak Gilbert (00:19:18):
    Pam, I had one question that came up that I think would be interesting to hit on here. So one of the questions was which teams or departments were the least enthusiastic about the AI tools? And the answer could be legal and regulatory, but I was curious,
    Pam Perino (00:19:32):
    And I wouldn't say least enthusiastic, but probably the most cautious, which is very to be effective and to be honest with you, didn't make me flinch at all. I work very closely with our regulatory manager, we have a very good relationship and I loop her in very early on any new things that we might be doing. And so initially I talked to her about these tools and talked to her about what we wanted to do and why we wanted to use them and things like that. And then she helped open the door with our legal teams. And also too, it was definitely a lot of back and forth before some of these contracts have been signed and just verifying that our review process, our approval process will remain the same that we have our product brand management team reviewing and approving. And then from there, our final approval for all of our content goes to our regulatory team because again, being in that food space, and I'm sure Todd can relate to being in a healthcare space somewhat, those things are so important for our consumers. We want to make sure that we're being very transparent and giving the consumers all the information they need to feel that they've got a trusted brand that they can enjoy. But definitely, definitely regulatory and legal. But again, it's all about building those relationships and bridges and just having those conversations early on and explaining to them the whys and the where fors and thinking about how you're going to work with these tools.
    Lauren Livak Gilbert (00:21:05):
    And it's about providing the context. I think you can treat AI in a similar way to how e-commerce was treated when it first became really popular. You need to provide context to the broader organization about why it's important, how it can help you. And I think that applies to every single function. So that makes a ton of sense where they need to understand the why around what is involved, what they need to think about, how this is going to change, how it's going to be perceived. So that makes a ton of sense.
    Pam Perino (00:21:31):
    It's like any new technology, it's new. Exactly. Scary. It's exciting. There's a lot of different perspectives on it, but it's here and there's great opportunities with it. It's just a matter of figuring out how best to use it for you and your company and the work that you're doing.
    Lauren Livak Gilbert (00:21:50):
    Perfect. Well thank you so much, Pam. We will hear more from you in the fireside chat, but at this point I want to hand it over to Todd. So Todd, do you want to kick us off here?
    Todd Hassenfelt (00:21:58):
    Yeah, no, thank you Pam. That was amazing. And like I said earlier, I'm at Colgate, Paul model, very talented company with a lot of great people before Colgate. Just to give some context, I have seen, I've been at two other big companies. I've been at two startups including a food one in Simple Mills and one probably in between where my e-comm journey started with glam bit performance nutrition. But I've been on sales, I've been on marketing brick and mortar. So I think all of these kind of come together now at a global role at Colgate with so many talented people everywhere is how do you help connect dots? And thankfully we are, we have a mindset of test and learns. So I think what I'm going to talk about to start off here is well, how do you bring that mentality, how do you bring it to life from a gen AI and from a content perspective?
    And first I would say is you have to first educate what are we trying to do here? You have to overcome, this is not just the shiny new toy, a baby shiny a baby new, but let's make sure it's a tool and not a toy. So what can it help us with? What can it help us solve that we haven't yet? Or what can it help us solve faster? But I think you have to do a baseline kind of education. Here is a use case, here's a specific things that we're thinking about, but as Lauren knows, this line is not about planning for perfection plan for pivots, not perfection. But if you can do those ahead of time, say, Hey, if this doesn't work, we're going to try this. If this doesn't work, we may go this angle. So I think that's one part, it's just kind of get everyone to start to be aware of what's going on and what you're thinking about doing.
    And then that obviously includes storytelling. If you go too technical, you're going to lose some people, you're going to lose a lot of people. If you don't go technical enough, your IT teams and those that are technical are going to think it's too loose. So how do you find that right? Storytelling again, how can you help others in life? In work, people will gravitate towards those who help them and help save them time. So how can you story tell with this situation? And really any right, as you start to do that, you should start to see, all right, who's asking some follow-up questions? Who's slacking me with some more information? Who's maybe joining a channel or whatever you have internally from Slack to Yammer to teams, whatever it may be, Google and who's kind of curious. And I think when you think about, alright, we want to test a provider, we want to test a process, really think about the teams and really it's not just team singular but teams because we will talk about next, it should be collaborative and cross-functional, but make sure you have those that are kind of interested like I mentioned, but also that have some time.
    No one's going to say, I have time just looking for things to do, but who is maybe staffed right in a full capacity, who has maybe a business need that would tie into content if we're talking about PDP content in this case. So really think about those teams and how can you connect them. And hopefully you have these relationships already and it's just connecting the dots. But I think if you pick the wrong team that may be interested but doesn't have time or may have time but isn't really focused on PDPs or whatever your tests may be, it can go sideways. That cross-functional piece, the collaboration, we always talk about silos and working together and yes, everyone wants to work together, but how do you really find the win-win situations for people? And that comes down to the storytelling. Don't assume the other departments work more than they do or vice versa, but how do you say, Hey, here's the gaps and where we are at now, what are your thoughts on this?
    Okay, now this team, so marketing gives or brand gives some perspective. Now we go to it, we work with legal, a lot of cross-functional work. That's hard. And I would say you need to do small group meetings before you have the big event meeting or whatever, a big event email. You need to do both. If you're only doing small group meetings without tying it together in a big event, then it's too loose and it's not really connecting the dots. But if you only do the big event or the big email, then you really don't get the adoption because there may not be full understanding, there may not be full acceptance of what's going on there. So cross-functional collaboration is key and it's difficult and time consuming, but it's worth it. And then I think the last thing is tying what you're doing to some kind of quantitative piece.
    And part of this, we just talked about storytelling, that's one side, but the other part is how can you tangibly show that you're making improvements that there is an impact here? And that doesn't always have to be for traditional dollar ROI obviously that gets the most attention, but it could be a time savings. It could be we're talking PDP content, how many pieces we've created. It could be the improvement of conversion scores depending on which retailer you're looking at that provides that data. It could be the impact of a lot of different factors around your digital shelf scorecard. But I think when you think about scorecarding, there's a quantitative piece and there's a qualitative aspect of it that visit or any other partner like that can help now really enhance your digital shelf. So on the next slide, let's kind of look at this a little bit closer is on the left side it's more the qualitative piece.
    How do you amplify image iterations based on the scoring that is now qualitative of the images? And we'll talk about that in a second and then later I'll talk about a digital shelf partner of ours and one of their new betas that does happen to be gen ai. Well, it's in here today. But really when you think about whoever your digital shelf partner may be, that is more a quantitative look always important. But combining the two now makes it really powerful. So if we go back to the left side and kind of the qualitative and thinking about, well, how do you do this in your organization? Everyone's different. Every department may be different regions, brands. So again, there's not going to be a cookie cutter way of doing this, but I think if you can take your images, your existing images that you have to start and think about the types of content that you have.
    Because when people say pd, well first of all, when they say content, depending on who you're talking to, that may mean social, that may be in PD DP, that may mean a blog. So define what we're talking about here, right? So for this purpose we're talking about PDP content, but take your existing and then within that PDP content, you probably have innovation situations. You probably have packaging refreshes or claims refreshes. You probably should if you don't yet have data optimizations. And that's where something like this can help. But how are you adjusting to what the data is telling you, what the consumers are telling you, what your retailers are telling. You probably have some retailer scorecards you're getting. So how do you really look at all this? And then also, how are you getting rid of cleaning out the old stuff so that you don't have a poor new person or new to brand come in and go to one SKU and see five different versions and not know which one's which.
    But using visit in this case, you can take your existing or even new potentially consideration, produce content and put it through their audience lenses, have benchmarks, scoring against as granular as you want to go, depending on your need. And really look at how are we scoring, how is this most likely to connect and convert our consumers? Now I think the muscle most CPGs got to think about when you think about PDPs is maybe you're doing it once a year, maybe twice a year, but usually it's if you have a brief or whatever process you have to get that content. It's like here you go, we're done. The change here is now with something like visit, you can do iterations and that's good, but that also creates more work initially to get that mindset change. But it gets you to the right answer quicker because if you can figure out what's not working and it doesn't go on one of your retailer sites for six months or 12 months, think about that conversion advantage that you have.
    So you kind of think about this flywheel here. You really want to build that process or enhance one if you have it already about creating, optimizing based on those kind of categories I mentioned score, iterate and keep the process going. There's no stop here. This is continuation. Then going back to the quantitative piece, ideally you have some kind of digital shelf scorecard going on. And now we're lucky enough just recently to be one of the closed beta partners with Ask Profitero, which is a generative AI tool, and there's really two portions of it. There's one which is content health. So looking at one of our markets and a couple of our retailers and looking at the copy portion, this is all copy now, not an image portion, like a visit that we just talked about, but are we infusing the keywords that we should do?
    We have optimized content. So there's a way of now prompting and looking and having suggested copy that could be considered to publish after it goes through human in the loop review, legal, all the processes that you may have brand. There's also a part that's an ASK analytics portion, which takes all of our existing digital shelf data. And now outside of content, so reins and reviews, looking at pricing and promotion, looking at availability, looking at retail search, how are we doing? Where are our opportunities? So think about this as the potential now to get to the answers faster for your practitioners or for those that are not hands on keyboard or as close to it, a different way to connect with them. So I think if you think about this and whatever tools you use or whatever process, there should be a qualitative and quantitative portion to this, but know there's going to be changes in ways of working.
    And so on the next slide for selling this into your organization, because this change management, no matter the topic is difficult. And I think especially when you're thinking about generative ai, you have to be balanced with the risk versus reward messaging. If you only come with only the rewards, the organization may think it's going to solve everything and then you, you're negating the risk, which is going to then have a lot of teams say, Nope, they're not thinking about this fully. If you're too much risk, you might stifle creativity or people experimenting in a responsible way. We know there's probably two ends of every organization, one that is going to wait for the checklist or the step-by-step guide of how to do this. They don't want to do anything wrong and that is fine. And then there's some that get a giant furniture set to put together and never look at the instructions and they just dive in.
    So both are fine as long as it's in a responsible way, but make sure your messaging is balanced. I think it's also kind of touched on before, encourage responsible hands-on experimentation with existing ops. We all have existing opportunities and things to solve. Let's not go chase ones that we haven't created yet because there's new tools, information is key and sharing is key to connect, but don't saturate 'em, right? And that may vary by team and by person. Don't overdo it with them. And then I think as you go through this journey, it's important to celebrate cross-functional successes and smart failures. It's not going to be perfect again, right plan for pivots, not perfection, but do that broadly and timely so that people, you build some momentum on both, here's what worked, here's what didn't, so don't try that, but here's what worked now, who can top it in a positive constructive way. So overall, I think it's really important to connect the dots and I'm lucky enough on our team at Colgate to have Elisa Levinsky who has just done a wonderful job connecting different teams and helping amplify the messages, find those type of people who really seek out and want to help others and help the organization. And I think you can find your balance between risk and reward.
    Lauren Livak Gilbert (00:34:49):
    Todd, the best wedding advice you gave me plan for pivots not perfection. So I apply that to many different things in my life. So I appreciate that. But Todd, thank you so much for sharing all of that. I actually have two questions that I think would be really relevant to what you talked about. So one of the questions that came up was around cross-functional alignment in the organization. So there seems to be a gap in identifying the correct AI tools to facilitate advancement. How does an organization vet the current avalanche of offerings and find out what works best with the least amount of trial and error? Do you have any thoughts on that?
    Todd Hassenfelt (00:35:25):
    Yeah, I mean it's tough. You have to be realistic in your expectations because you can't say whoever you decide. And that's a whole different thing. Who is going to be reviewing these, right? And saying yes or no or maybe, but there's too many to say. We're going to go through all of them realistically. So look, well one, again, looking at those use cases, what are the best ones that could help you, your company, look at the terms of service agreements work very closely. You should be working closely anyways with your legal teams and IT teams. But if you're not, you start now and really work with them on that portion to evaluate it. Work with your security team to look at this because there are varying levels on the other side of these gen AI tools of a lot of people or a couple people on the other side.
    And do they have agreements that are favorable or agreeable to your particular company? It's going to vary by each, but I think you can also look at, well, where are the needs? But maybe also even if you have 'em blocked, are people across different divisions or different regions trying to log into certain tools and maybe track those and say, oh, okay, these five are getting the most hits. We're blocking 'em right now just out of concern. But maybe those are the five that we start doing the security review, the legal review, the use case review, and then maybe ideally you're educating more so if that's starting at a central kind of team across a collaborative team. But ideally you are sharing that knowledge so that it expands if you're starting in a global piece, but then go to different regions, geographies and teach the local teams so that they can do it maybe. And now you can expand the number of tools that can be reviewed, but then they're sharing their learnings across the global with the global team to help connect those dots. But set expectations, you're not going to review all of them.
    Lauren Livak Gilbert (00:37:23):
    Yes, prioritization. And
    Todd Hassenfelt (00:37:24):
    Just because you review it once. Yeah, just because you review at once, these change, the terms of service agreements change, you have to keep on it. It's not either you rejected it or accepted it or said we're going to try it. You have to keep active on those.
    Lauren Livak Gilbert (00:37:39):
    Yeah, it's an evolving space. So I think that's a really great call out. And then one other question before we get to the fireside chat, and I'd love Todd and Pam, both of you to answer this because we haven't gotten to this yet, but a big concern of AI for some people is will this eliminate my job? Will it make me not needed? So one of the questions was around have you seen adding AI to free up time for people on your team and have upper management think about either downsizing or eliminating a role or changing the team structure? Can you just address that? Because I think that is a concern for some people when we talk about ai. Maybe Todd, I'll start with you and then go to Pam.
    Todd Hassenfelt (00:38:16):
    Sure, sure. I'll just say I think in general I get the worry about that, but we just talked about earlier, everyone is so busy and there's a lot of stuff you're not doing that you want to. How does it make what you're doing faster? And for example, if you're not a finance person, you're not going to be doing the gen AI kind of spreadsheet stuff. If you're not a creative person, you're probably not creating images necessarily. So how can you use it more to be more efficient? Because there is still a lot of work to do. Every company is going to make their own decisions. But I think if you use this more as a tool to be more efficient, to look at data faster, to get to the answers faster, to illustrate things faster, that should help your role and more worry about how to learn it versus how to avoid it or hope it doesn't happen.
    Lauren Livak Gilbert (00:39:09):
    Thanks Todd. Pam, what do you think?
    Pam Perino (00:39:10):
    Yeah, absolutely a hundred percent. I mean I haven't felt that or really gotten that at all from maybe in our organization. And I think it's really, again, it's a tool to help support and us be better and faster at what we're doing. And at the end of the day, we still need that human eye, that human intervention. There's no way that, again, would've never been able to use some of these tools without getting those regulatory and legal buy-in like, okay, our process, the review process is not changing. We're not just going to turn this on and have it spit out things and off we go. It's like you can't can't do that. So you still need humans. It's just a way for us to be more nimble and fast and be able to do things quicker and better. And the digital shelf is such a fast moving space as we all well know. So there's definitely, I don't feel that that's a cause for concern.
    Lauren Livak Gilbert (00:40:09):
    It's a big focus on efficiency. How can we do what we're doing more efficiently and use it to help enable our role. So I'm glad we addressed that though because I know there's definitely some concerns around that in the space. Now one of my favorite parts of the webinar, the fireside chat. So I know there's a lot of great questions. We are going to answer some of them in this fireside chat. And then we will definitely get to the additional questions in the q and a. So we have a bunch of questions. I'm going to throw them to either Todd or Pam and we can jump through them. So why don't we start with Pam. What was the biggest hurdle in bringing AI into your digital strategy overall?
    Pam Perino (00:40:45):
    I think probably just getting management to understand what the tools are all about and what they can do and why we want to use them. And then just to reiterate again, as far as getting legal and regulatory on board for them to review and look at things and how are these tools going to be used and what guardrails are we going to have? Are there processes that are going to need to change? Things like that. And that was probably our biggest hurdle. I didn't feel like we had any pushback or concerns from brand marketing or our creative team or things like that. I think it's just educating also too and talking to them about what the tools can do and how they can help all of us. And I think it's really about communication and really building on the information and making sure that you're really giving people an opportunity to understand what these tools can do to help support their work as well.
    Lauren Livak Gilbert (00:41:43):
    Pam, a double click on that. Did you have any challenge weaving it into your overarching process? So a question that came up was any learnings you have from that? Because it's not just working with legal or regulatory, it's marketing, it's sales, it's the broader process.
    Pam Perino (00:41:59):
    And I think also too for our creative team, it was asking them to perhaps put another step in our image creation. It's really using as far as for visit, using that tool as we're creating images and let's have the tool open, let's be creating, let's measure the images, which one do we like, which one's got the best score? Okay, let's go with that one. And really learning the tool. You don't want to create more work for somebody, but you want them to understand how it can make us all better in what we do. So I think that was probably one of the harder challenges is just everybody's so busy and has such limited time. So how do you get them to want to work with these new tools and get them trained and get the buy-in and have them see what the tools can do. So that's probably one of the things that we had to get over a little bit, get over a hurdle.
    Lauren Livak Gilbert (00:42:53):
    Perfect. Todd, what about you?
    Todd Hassenfelt (00:42:55):
    I think from a gen AI perspective, this is a great one. If you look at any kind of change management, it's like five A's right? It's one is awareness. How do you get the education to hear about it or at least have an opinion, good or bad? So you have that awareness piece. Then you have to get to acceptance. And these all happen in different stages. So then you get people or teams to believe, well, this could work. I think it will work, it could, okay, now if you really get them believing that, then they adopt it. So I'm actively engaged in using the tool or talking about the topic. So that's great. So you have awareness, acceptance, adoption, then it's advocacy. Now you're building advocates and you have others helping say, I tried this, look at the images that we scored here and how much faster we got to these answers.
    And then I think there's also adjustment, right? Don't just rest on your laurels, what you learned last year and a little bit or last month depending on the topic here. But I think if you kind of think about those five stages in this and really any kind of change management, bring it in and know that you're in different stages with each. But as people move to adoption and advocacy, help them with the awareness and the acceptance in other divisions, other teams, I think you do have to address the concerns of, well, what does this mean? How much time do I have to invest? I think that's a key point is the initial investment in learning, but it's going to get better than afterwards. And that's a tough one without seeing it and actually happen in real life for them personally.
    Lauren Livak Gilbert (00:44:35):
    And Todd and Pam, actually, we'll start with Todd. How long did it take to get your first AI initiative off the ground? Was it weeks, was it months? Or you can just give me a range so people understand how long it took internally to really get that off the ground.
    Todd Hassenfelt (00:44:51):
    Well, I mean it varies, right? Because there's different initiatives going on. I mean, if you think, I guess of the very first one, there's probably weeks in terms of use cases and those people that are the ones that get a furniture set and don't look at the instructions, but done in a responsible way of guardrails that we have. But I think it's then okay, in each situation, no matter how long it took, is it scalable? Yes or no? Is it going to be helpful? But also then how do we look at these tools in terms of what can they bring to us? Is this the whole buy versus build conversation too? This applies to this as well, but I think it's an iteration and it depends on by department and by use case where it is. But you can have some in weeks, you can have a couple in days if you're doing competitor research on publicly out there, you can do summarizations of PDFs of let's say an earnings call that's public information and have a summary. And maybe that's a different kind of competitive assessment than you've done in the past. You can do images scoring with visit very quickly, but again, there's a level of it shouldn't be looked at in weeks or months or days or time. Really. It should be looked at how are we helping, and this is continuous. We're really not done with this topic and it evolves all the time. So I wouldn't put too much time on it, maybe just to get started, but really know that this is going to be a continuous cycle.
    Lauren Livak Gilbert (00:46:25):
    Thanks, Todd. Pam, how about you? From your side, in terms of time to get things off the ground?
    Pam Perino (00:46:31):
    It varies. I think it can be at the minimum a few weeks to a couple months perhaps. And again, it kind of depends on I think the scope of the tool and what you're doing and things like that. But it's definitely a process. And I know just from these tools, it's a matter of, again, kind of getting everybody on board. And then of course getting things through legal sometimes can take time as well too. So it just varies. I don't think most companies can do something in this in
    Lauren Livak Gilbert (00:47:08):
    Days or hours. That's the ideal world, right? Yeah, exactly. One day maybe. Yes. Alright, so we've talked about this a bit, but I know there's a lot of questions here around the story that you told to your executives and how you brought it through with the leadership. So how did you bring your organization and leadership on a journey with you as you went through this process? And a lot of the questions that are in the chat here are around quantifying it and having specific data to showcase that it is valuable. So maybe we'll start with Todd. What was your kind of journey and do you have any of those quantifiable metrics that you were able to share in that story?
    Todd Hassenfelt (00:47:51):
    Yeah, I mean it goes back to one of the points I talked about is especially when talking to senior leaders is truly that balance of risk versus reward. So you could show thinking early on as this was developing and their stories every day you have to curate which stories you show and all that. But from a reward perspective, it is kind of thinking about, okay, well how could we apply this by which department, which should we go test out? And I think there was that initially of, okay, here's a few departments, here's a few use cases that we could use. Here's some tools that we think would be the best ones to start with. But also the risk portion of it is, and I talked about those terms of service agreements, looking at what levels of protection do we have from each of them, what do we have to be cautious of?
    How do we make sure teams are educated? And so I guess when you think about talking to the senior leaders about this, it is just making sure that you are trying to put as much out there, again without saturating, but really be specific on which use cases could work and have some examples ready for them, whether it is around PDPs, how it could look at innovation. I mean, this is where the storytelling comes in and use the word potentially a lot because it's not for sure, but this is where we could be. I think also one of the things that's key, and this is really with any conversation inside organizations, but very early on, there were already public examples out there from other CPGs and other companies in general of what they were doing with generative ai. And I think that type of comparison slash competition to share is sometimes helpful that it is that balance of not being maybe a trailblazer in the space, but also not falling behind and kind of that fast follower mode depending on what use case you're looking at.
    Lauren Livak Gilbert (00:49:57):
    Thank you, Todd. Pam, how about from your side?
    Pam Perino (00:50:00):
    I think definitely my manager has been a big supporter and proponent of these tools and has entrusted me to do the due diligence and all of that to work with them and things like that. So I think it's kind of begins with I think your manager direct leadership and kind of showing them again what these tools can do and having them really have a understanding of the efficiencies they can bring and what the tools can do for us as far as the organization and making us better at e-commerce. So I think again, and as Todd mentioned, it's like you want to make sure you're giving the right information to the right group and what impact or what are you solving for them? It's a tool that we're using on our digital team, but how does this help support them? And I think that's also part of it as well, is realizing that these tools can also benefit other parts of our organization. And I think that's also part of how you bring somebody on a journey with you and get them to believe in a platform or a tool and want to use it as well and engage with it.
    Todd Hassenfelt (00:51:21):
    And Lauren, I think now that we're kind of further along, if you think about doing this in first quarter versus now, now there's also more examples of how the retailers are trying to embed a generative IP. So okay, now how are we're going to partner with them, but then also the consumers, and again, there's different sides. Some are still wary of it, but some are embracing it and some are saying, yeah, I'm okay with this. So you really have to monitor those type of surveys, data points that come out from what your retailers are doing and how are the consumers looking at this and what is the overall impact?
    Lauren Livak Gilbert (00:51:54):
    A hundred percent agree. And I think it is having a pulse on what's happening to that point, but also having the pulse on that change management curve within your organization and how you're bringing people along that journey. And Todd, I liked those five A's, what were they again? Awareness.
    Todd Hassenfelt (00:52:09):
    Yeah, so you have to have awareness, then acceptance, then adoption, then advocacy, and then adjustment. Don't get stuck. Don't get stuck on what you learned. Things are always changing. But I think those five can apply to really most change management situations and you're going to be in all five of 'em at some point depending on the scope.
    Lauren Livak Gilbert (00:52:29):
    Change management is always a critical factor in anything that you're doing. So definitely keeping that top of mind on the AI piece as well. Okay. So if we look into the future, if we had a crystal ball for next year, how do you plan to continue leveraging AI in 2024? Pam, maybe we'll start with you on that one.
    Pam Perino (00:52:46):
    Sure. I mean, as I mentioned earlier, we've already incorporated visit. We started working with 'EM earlier this year on measuring our images and looking for new design. For right now we're working on new designs for our neuro images, and then we're going to be rolling out some new carousels and new design techniques that we want to do there. But I think one of the things that we're just starting to work on again, as I mentioned earlier, is with the content creation, marketing copy creation with citation and rough draft pro and being able to really start to work with them, use that tool and how do we incorporate our brand voice and our brand tonality and things like that. And also too, how do we really hone in on content for our retailers specifically what their needs are, what their restrictions are, things like that. And so I think these tools can really help us do that, but that's probably our next kind of thing over the hill right now is probably the marketing copy and we're just kind of getting started on that journey.
    Lauren Livak Gilbert (00:53:59):
    Many opportunities in 2024. Todd, how about you?
    Todd Hassenfelt (00:54:03):
    I think let's just hone in on PDP content. This is big enough, but it comes down to ways of working because there's multiple points of failure. So let's say you ramp up just the creation portion of it or the optimization, and now you're creating thousands of images or copy bullet points and titles a month. And in the past, that would've taken two years. If you're not ramping up the moderation, the review of it, the assessment from a legal regulatory brand perspective, that's going to create a bottleneck. And it's no one's fault, it's everyone's fault because we're all in it together. But I think it's a ways of working of making sure you balance that you don't over resource one side or even if you over resource, then all the reviewers of content let's say. But then you don't have content coming in now you have a lot of headcounts sitting around going, where is it?
    But also even just the creation, right? And the iterations that I mentioned that you can do with visit, and now you can get to that answer faster of what's working or what's not. And then thinking about your PM DAM solutions, thinking about, okay, is it getting to the retailers? Now, I may really have that opportunity to vary my content by retailer more often and how do I connect with my retailer and negotiate with them if we're all helping each other on conversion? But ideally, if we're doing all these things and that ways of working gets figured out and there's a lot to figure out there, we are connecting and converting with our consumers in a more relevant way, that should help a whole bunch of different measures. But I think this is a balance, right? This should not be just because this is digital commerce or e-commerce. This is not just conversion. This is also if you still believe in funnels, top of funnel, bottom of funnel, we all need to come together and figure out how do each of these little parts, we're not replacing whole things, we're just making things a little bit better or maybe a lot better depending on the situation.
    Lauren Livak Gilbert (00:56:10):
    I love that. I think that's a great point, Todd. We're not replacing things just as both of you shared around what AI is doing, it's supplementing and that's really the main focus. And a majority of that supplementation is around efficiency and scale. And that's really what it enables a lot of digital teams to do, because we are not mighty teams of 30, 40 people. We wish we were, but we are not. So how can we mimic that kind of work? And AI is a huge enabler of that. Okay, last but not least in terms of our fireside chat questions, and then we can hopefully get to some other ones. There's a lot of questions around measuring the success of AI and specifically around can you connect it directly to growth? Have you connected it to conversion? How are you thinking about the specific metrics that you're using to measure the success of AI and also the success of the tools that you're using? So maybe Todd, I'll start with you on that one.
    Todd Hassenfelt (00:57:02):
    Yeah, I think it's a way of enhancing, assuming you have digital shelf scorecards, right? And you've been showing the progress over time of those, and again, those quantitative measures, now you potentially with visitor or another tool that looks at it from a qualitative perspective, you could show how many pieces of content now have gone from this average score to the higher score. So you could look at it that way for the retailers that you do get conversion data from. You could do some AB tests or do some testing here, but you could see is it an improving conversion, which should then improve sales. But I think about when you think about PDPs and if you haven't bought into digitally influence sales yet, you should. And the DSI has done a lot of great research on that. And a shout out to Lauren, she did a great digital deep dive podcast with Erin Conant recently.
    So everyone should go listen to that on all things organization and processes. Thanks Todd. But I think of course, no, it was fantastic. But when you think about PDPs, it is not just that whatever 5% impact, right? There's digitally influence sales, but now think about your social, your search, like your email copies, potentially all that money that you're funneling in different ways, a lot of times leads to a link that's a PDP. And for that new to brand or new to new product type thing, you lose them if that PDP is confusing them versus connecting and converting. And I think also the impact of in-store retail media, and yes, there's been screens in stores before, but how they're being used now is changing and there's already pictures out there. Still smaller scale, but it's going to get bigger of those PDPs are going to be on those big screens. However, they are smart carts, end caps, coolers, whatever, as well as their phone. With in-store apps, there's going to be a whole different level of scrutiny of the quality of PDP that maybe wasn't there before. And so get ready for it. And I think generative AI and tools that measure the qualitative impact can get you there before you have to be there.
    Lauren Livak Gilbert (00:59:10):
    I love it. Pam, how about you?
    Pam Perino (00:59:13):
    Yeah, I was going to say we're still fairly early in this journey with ai. So we're definitely at the current tools that we have in our organization for our digital shelf measurement in regards to conversion and things like that. And what capabilities do they have for measuring success and conversion, things like that. Are we seeing an improvement in sales? Are we seeing better click-through rates? Things like that. So I think we're still kind of moving forward with that. I don't have any specifics yet, but it's definitely something that's obviously very important to us because if we're investing in these tools, then we want to see, okay, if we are making decisions and the tools are providing this information, how is it better than what we were doing or are doing? And so again, it's like any investment that you make. So we're still a little early in that journey, but we have a lot of things in motion right now.
    Lauren Livak Gilbert (01:00:15):
    Perfect example of a test and learn, right? So figuring out, seeing how it can drive impact, measuring that and then moving forward to potentially scale it more. So I love that a lot of what we talked about is using the basics and the foundations of implementing e-commerce and any other project within the organization with some slight tweaks because it's a different technology, but a huge thank you Pam and Todd, thank you so, so much for sharing all of your insights. Eli, thank you for setting the stage for us around AI and around the visit capabilities.
    Peter Crosby (01:00:45):
    Thanks to all our guests for allowing us to repurpose their digital shelf leadership stories on the podcast. Become a member of digital shelf to keep on top of best practices research and more thanks for being part of our community.