Leveling Up Your Teams and Customer Experiences with AI | Wunder 2023

Media Thumbnail
00:00
00:00
1x
  • 0.5
  • 1
  • 1.25
  • 1.5
  • 1.75
  • 2
This is a podcast episode titled, Leveling Up Your Teams and Customer Experiences with AI | Wunder 2023. The summary for this episode is:
Introductions
03:57 MIN
How Can Marketers Prioritize the Most Impactful AI Driven Opportunities?
01:44 MIN
Implementing and Approaching Applications of AI
00:58 MIN
How Can AI Be Harnessed to Deliver Hyper Personalized Experiences to Customers?
04:24 MIN
Effective AI Driven Strategies to Deliver a Personalized Customer Experience
04:42 MIN
Which Industry Has the Most to Gain From AI?
00:39 MIN
How Can Companies Start to Upskill Their Marketing Team to Work Effectively With AI?
04:26 MIN
How to Strike the Balance Between Automation and Creativity
01:18 MIN
How Can AI Be Utilized to Create Engaging Content for Your Audience?
04:07 MIN
How Can Brands Ensure They're Using AI in an Ethical Way?
04:36 MIN
How to Stay up to Date on Artificial Intelligence
03:00 MIN

Megan Kresinske: What's up, Wunder? How's everyone doing? Come on, a little better than that. First, I want to apologize to my panel for bringing you on for such an unpopular topic. So thank you guys so much for agreeing to chat with us today. No, all jokes aside. I'm Megan Kresinske, I lead the product marketing team here at Wunderkind. I'm really excited to talk with this super powerful AI panel today since AI is obviously such a hot topic. I read an article the other day that says you can just slap the AI label on anything you see now and it will be relevant. In fact, CB Insights just says that there's currently 350 plus generative AI companies out there right now across 50 different categories. So a lot of buzz in the AI world. But they said retail is the sector that has the potential for the most adoption with AI. And I think since we're in a room full of retailers, this is very relevant for you all. So excited to discuss with you all today. So recognizing this is a hot topic and the countdown is on, we should dive in. But before we do, I want to give our panelists a chance to introduce themselves to you all. So Abran kick us off.

Abran: Thank you, Megan. We really appreciate it, again. So this is my I told you so tour. I was here last year pre ChatGPT, and I was like, " It's coming," and here we are, right? The little things showed up. So for those of you who weren't here last year, I'm the co- founder of Create Labs. I'm also an ambassador for OpenAI small company that created this little thing called ChatGPT. And also the creator of C. L.Ai.R.A, who I brought with me last year on stage, a virtual AI woman, a female influencer, woman of color. And now we travel the world. I recently signed with profits of AI, the AI speaking agency. So this is one of my final domestic stops. And then we're going to Prague and Qatar next.

Megan Kresinske: Nice.

Abran: Hitting the road because of this AI buzz.

Megan Kresinske: Love it. Love it. Gareth.

Gareth: Hi, I'm the British contingent. I'm Gareth representing, I think maybe atypically representing the UK. I'm the Chief Digital Officer at Kurt Geiger. Anybody heard of Kurt Geiger? Not enough, yes. So my marketing isn't doing that well. We sell women's handbags and shoes and our core principles really ring true to the sort of conference topics that we've talked about this morning. Kind by design is our strap line. We're all about inclusivity, diversity, color. My role, I'm a shopkeeper, so my AI background really comes from really practical applications of that for retailers.

Gary Pak: Hello everyone, I'm Gary Pak. I want to apologize in advance, I'm getting over a cold so my voice is a little bit raspy, so hopefully I can make it through the entire session. But I'm the president of CSC Generation. And what CSC is, it's an investment holding company that's been fortunate enough to acquire a number of well- known brands over the course of the last few years. And so within our portfolio we have companies like Sur La Table, One Kings Lane, Touch of Modern, Z Gallerie, as well as a number of B2B furniture wholesalers. And so think about CSC as a private equity company with a little bit of a twist where not only do we provide funding capital for these organizations to continue to operate. But we also provide shared services in the form of R& D specifically around technology, AI and some innovative things that we could do to really grow these businesses in scale.

Megan Kresinske: Cool, thanks guys. I think Gareth, I'm going to start with you. So I mentioned earlier AI is everywhere with this rapid emergence of AI applications, how can marketers, aka the folks in this room prioritize the most impactful AI driven opportunities for their specific brands or industries?

Gareth: Big question. My approach has always been about breaking down barriers to having good conversations at board level about these things. So although as a brand has predicted last year that everybody's now talking about AI, the actual understanding in the boardroom is very little about it. It might be from a movie they've watched or a conference they've been to. And every SaaS company that comes to you, every marketing, every MarTech is always saying they've got AI technology that's going to double your ROI, single line of code and you're away. inaudible can do that by the way. So I think my advice first and from my point of view as sort of a techie geek would be as a marketer, I do the CMO role and the C DO role for Kurt Geiger. So I place all that marketing and a lot of what we do is just try and get more and more board buy- in into what it actually means. And split down beyond just this umbrella term of AI and understand is it language models, is it machine learning, is it data analytics, is it predictive modeling? So that you have a more open and honest dialogue with the board to get buy- in. And it's not just the tech guy who's saying we should have AI for no reason, but there's an understanding of what this capability can enable you to do against your business strategy and against what your customer demand would be.

Megan Kresinske: Love that. Abran, I'm going to phrase this to you just slightly different as another techie AI geek, right?

Abran: Yes.

Megan Kresinske: What are your general philosophies when it comes to implementing and approaching applications of AI?

Abran: Well, we heard some great conversations earlier from folks that are in the accessibility and inclusion space and marketers in the room that have to be careful with how this is implemented to the public. And also internally we've heard of things like hallucinations, we talked about that yesterday. These systems can bullshit because it's a people pleaser. It wants to give you an answer whether it's right or wrong. So not only how you design it, but programming it to what not to say. Putting those right guardrails in there and adding additional layers that are just safety checks are extremely important when you implement these technologies. It's doable, so don't run from it. I would say embrace it, but just do it the right way.

Megan Kresinske: Yep. I love that. That's actually leads very nicely into my next question for Gary. So something that is near and dear to Wunderkind is delivering personalized marketing experiences. I think everyone in this room can relate to that. However, there are already concerns and debate that AI may cross into intrusive territory when it comes to delivering personalized digital experiences. How do you think AI can be harnessed to deliver hyper personalized experiences to customers without crossing the line?

Gary Pak: Yeah, I think the first step is just transparency with the consumers around how you're applying AI to benefit them and the customer experience that they're enjoying. There's a number of studies from Forester and McKinsey that have shown that retailers and implementing AI have seen anywhere between 20 to 40% improvement depending on the scope of AI that they've implemented. And there was a subsequent study that essentially said or asked to consumers, " Will you be okay with machine learning or AI algorithms to provide you personalization?" And the overwhelming majority of them said it was okay as long as their data was secure and protected alongside that context. So not violating their trust. So I think transparency and then also putting the guardrails that Abran mentioned around making sure that the AI just doesn't go batch shit crazy and doing a bunch of stuff that it shouldn't be doing. So there needs to be a certain level of governance that sits upon any implication of AI as well.

Megan Kresinske: Yep. Gareth and Abran, any thoughts?

Abran: It's scary what it can... I brought up this scenario yesterday. We were on the drive over here where an ad can be generated in real time on the spot as you visit a page. So it could identify who's visiting this website page, their demographic information and generated an ad customized to that person on the spot. Instead of having pre banked ads that you're trying to match to a person, you could just generate it on the spot content ad read, voiceover, the whole nine.

Megan Kresinske: Wild, it's crazy.

Gareth: I think there's an interesting dynamic going on here about privacy isn't there? Because quite right, when you ask people do they want more relevancy, they always say yes. My wife loves the fact that Facebook's listening to her all the time.

Abran: My wife inaudible.

Gareth: Because she wants to see the ads that are relevant to her and she actually thinks that's a great thing. I'm on the other camp, I'm like, everything's shut off. I just get BMWs and typical guy kind of ads to me and I know I've beaten it because there's nothing specific. But I think-

Megan Kresinske: I don't think that's how it works, but continue.

Gareth: It's the way I like it. I think there is a data exchange that people want, but a lot of the time with these sorts of technologies, people are experiencing them a lot before they're spoken to about it. So before we started talking about data privacy and data security, they were already leaking data out right, left and center and getting good and bad outcomes from that. It's our job, I guess as leaders in the field to make the bad outcomes less and try and make the education of understanding why there's a fair data exchange for those things, be positive for the user rather than that. And particularly as a retailer, we've heard about this morning, your brand equity relies on these things. You don't want to get caught out like you would do with diversity in a data breach or a bad use of people's data. You want to have this relationship. And so really it should be a core conversation in the business that says, " How do we make sure the outcomes that we are using this for are transparent perhaps, but in the right language." Because if you start being too transparent, people get freaked out. Because they're like, " You're doing what?" And you're like, " Yeah, but you're doing that all the time." So I think there's a way of bringing your consumer along. Depends on what you're selling, right? It depends on who your target market is about how developed their attitude towards this sort of language can be. Yeah,

Megan Kresinske: I totally agree. We actually were just joking around before this panel that I was on a flight previously where the captain was like, " Oh, we're not going to take off for a bit because the breaks don't work." And I'm like, "Huh?" I mean there's nicety to transparency, but maybe you could have just said, " We're having a slight issue, we'll move along soon." So I think it applies across the board. Okay, next question. Something I'm excited about is AI's potential to revolutionize the customer service experience. I was talking to Gareth and I know I'm the moderator and it's not my moment, but I did have a funny story to share. So the other day I was on hold, I won't say the company, but I was trying to upgrade my internet service. I waited an hour, finally got someone, and then they said, " Oh, Ms. Kresinske, we don't service your area." And I was like, " How is that possible? I have your internet." This is really shitty customer experience. And I waited an hour to be told something that I'm already a customer of. But anyway, it was a funny experience where I was like, " Well, if AI was involved, maybe I could have just bypassed this really terrible experience at least a lot faster." I at least could have had someone tell me in real time like, " Hey, you're not a customer." When in fact I am, but I am not any more because of that. I think there's a lot of benefit to AI and customer service. So I guess in your experience, I'll start with Gary. What are some of the most effective AI driven strategies to delivering a personalized and seamless customer experience?

Gary Pak: Yeah. You hit the nail on the head. I mean, I think one of the more frustrating portions of any business is the customer service contact center. Most businesses considered the call center, and so they're reducing costs as much as possible. Which means less call center agents, less customer service agents, longer wait times. And then oftentimes the systems are so disconnected and broken that you can't really get the answer that you're looking for or the resolution to the problem that you're trying to solve. And so I think AI, there's a lot of applications. You think about retail in particular, a lot of the call volume that comes in is WISMO calls, " Where's my order? When's it coming to me?" Those type of questions. And AI is the perfect vehicle to be able to answer that really quickly to consumers in a very safe manner. I think you have to shape it a little bit differently where it's not your standard automated voice response system because those could be a little bit frustrating as well, and you get lost in that as a consumer. But with AI technology evolving, there's opportunities where you could synthesize a real person's voice and then apply that so that the end result feels the same. You talk to a human, you talk to a bot, there's no real difference except the speed of service. It's accelerated and you're getting the exact answer that you wanted as well as part of the phone calls. So I think that's one of the main drivers of AI that we're applying within the CSC portfolio of companies, and we've had a lot of success with that.

Megan Kresinske: Abran, any perspectives since you work across a lot of different customer service industries?

Abran: Shameless plug, this is what I do as a business. So a lot of companies come to us and say, " Hey, we don't want to risk doing this ourselves. Can you help us implement our technology?" Usually the first discovery call, we discover 10 to 20 different ways that they use AI, and then we find one that really makes an impact. Two very interesting ones that are applying here. So one is a real estate development, they develop these communities all around the country and then they develop this high- end luxury building. They asked us to develop a C. L.Ai.R.A, like a personal concierge in the suite. So you interact with in the panel and can order Instacart for you, but also memory retention. So if you allow that data, then the memory retention can come into play of they learn you, they know you. They're like, " Oh, how was the little league game this week? Oh, you ordered this. You're running low on orange juice. You want me to order you more?" And have that interaction. Another one is for a university hospital here in Texas, shout out to Texas Tech. They are developing a group NDA maybe, kind of like this doesn't leave the room. So it's reverse.

Megan Kresinske: Noted.

Abran: Yeah, it's the reverse where they're doing better to engage in person. So what I learned from a doctor visit is that of a 30- minute visit, you see the doctor for five minutes because they spend 25 minutes transcribing, taking notes, coming up with the treatment plan, billing codes and all that. So we're developing an AI that does all that in real time. It's listening to the whole patient visit, doing all that in real time so that the doctor can gain 30 minutes of engaging in real life with the patient that they don't have right now. So there's a lot of... I like being industry agnostic and seeing every industry use case because I learned so much. And sometimes we will build something for marketing or fashion that applies to education or vice versa.

Megan Kresinske: And I recognize we're in a room of brands, but I'm curious, Abran, because you work across so many different industries, which one do you think has the most potential to benefit from AI?

Abran: Education. They're a little scared of it right now, so a lot of my tour is speaking to education audiences, helping them embrace it. Speaking to leadership, policymakers, going to DC and speaking to Cardona who's the secretary of education on the 20th. So in a couple of weeks. And I think it could do a lot for personalizing education, but they have to get over the fear first and find a way to do it the right way.

Megan Kresinske: Of course. Yeah, okay. Moving on, Gareth, we're going to talk about navigating the AI talent gap. So the demand for AI experience is high, but as of right now, the talent pool is a bit limited, there's not a ton of people. There's new jobs coming up all the time, prompt engineers and so on. But how can companies today address not only the talent gap but also upskill their marketing teams to work effectively with AI technologies?

Gareth: That's a big question. I think that there's two routes, isn't there? With these things, as I see it, there's lots of people who, like myself included, are plugging their own product or their own services here with room full of super and smart people who hopefully will take on some of those services and some of those products. And there's always the balance that we have between getting the right advice from the right talent, especially our organization is like a medium- sized business. So we've got 200 people in head office. I've got 35 software engineers that work on my team and then augmented with teams outside of that. And it's hard to have the budget to have real talent in these fields in- house from the start. So I think there's two ways to approach it. The way that we've approached it at Kurt Geiger certainly is look for the experts in the field who can help you of where you are on the journey, give practical things that you can do and then start the conversation internally. Because a lot of the time, like I said before, a lot of the engineers that work for me are using AI all the time. But they don't say they're an AI specialist, they're not going to be setting my AI strategy. In fact, I don't believe there is an AI strategy for our business. It's more about where it will just morph into everything that we're talking about.

Gary Pak: It's an efficiency.

Gareth: It's absolutely, exactly what I was going to say. Thank you. So the CFO loves it, right. So the CFO is like your contact center, it's a cost center to him. To the marketeer it's an opportunity to have a dialogue with the consumer. So we're using things like in bots responses and canned responses for real people. We're using AI to make them more friendly or make them more relevant to a different community. And so there's I think steps along the way that you can do that. So it's not necessarily about, " Okay, I've got to have an AI department in this..." Because we're a retailer, so we are the creative people, we sell handbags and shoes, so most of people in our business create great handbags and shoes. So I think a bit of both. I think breaking down that and making people understand that it is applicable to their job. And it's not this big AI over there with the CIO, is part of everybody's job, and then getting some advice for where you are on the journey. Every retailer is different and every board is in a different place. So some will be really embracing of a total education piece by somebody really smart. Others just want to be slowly educated on this with real use cases. So I think a balance of the two would be my approach.

Megan Kresinske: I love that. And Gary, you are investing in...

Gary Pak: Yeah, I think AI's kind of like a shiny new toy, and so everyone's running to it. And OpenAI, there's some parallels between what OpenAI is doing to what Apple did in 2008 with the iPhone release, where it's creating kind of this sub economy. And you're going to see... You mentioned all these startups leveraging AI to build SaaS platforms that are going to empower businesses like that we have here. I think for a business to make it work, you have to really focus in on the few areas where you think it's applicable to your business, where you get some sort of ROI. Whether that's revenue growth, whether that's cost savings, whether it's scalability of work efficiency, whatever it may be. But there has to be target goals that you put in place and say, " Does AI fit this goal? And can I accelerate hitting these goals by leveraging AI." And then going out and selecting a firm like operations, right firm to come in and help out and accelerate that? Or if you want to make a decision to build your own internal teams and make that commitment, then that's up to your own choice, but it's going to be heavy. The cost of engineering is really expensive, and so I recommend kind of dipping your toes and experimenting, see what works. And then as you experiment and iterate and you land on something where there's a tremendous amount of ROI, that's when you start building infrastructure internally to help support that.

Megan Kresinske: Great. We're going to move on to balancing automation and creativity since we're in a room full of marketers. So obviously we've talked a lot about today how automation can free up a lot of valuable time for marketing teams. But really how do you strike that right balance between automating mundane tasks and fostering creativity? I think Abran, you have a good point of view on this one.

Abran: Yeah, I think the Coca- Cola case study was a good one where they produced an amazing commercial with AI. It was a great demonstration of how AI could be used in an ad. However, there was no labor lost including AI. There was still a video production team. There was still the editors, there was still the writers, and it was just added to it, not replacing anybody. If you ever find it on LinkedIn, they have a good behind the scenes of how it was produced, and it was just a supplemental piece to the whole creative workflow. But I would say give your creative team an opportunity to learn those tools. It's just another tool in the toolkit. Don't do anything rash. We're like, " Oh, we have this license to this AI toolkit. We don't need this creative person or this department." That's being rash and they're the best ones to use the tool.

Megan Kresinske: Yep, agreed. I think this next question relates to everyone and to this particular topic is around AI and content creation. I think most people I've spoken to recently and into this room, that's probably the number one area they're starting to adopt AI, is to help scale content creation. My biggest concern is we start to live in a world of content saturation where everyone's saying basically the same thing. So what are some innovative ways AI can be utilized to create compelling and engaging content that resonates with your target audience that doesn't get lost in a sea of AI, for lack of a better word? Gareth, you want to kick us off?

Gareth: I think we're already there. I think the over proliferation of content is already there. I think it's filtering it now and personalizing that experience for people, which is key. I think somebody told me this sort of analogy, and I really liked it, that people are a bit scared of the AI thing destroying creativity. Is a bit like mathematicians when the calculator came in thinking that people weren't going to be great at maths anymore. But what it actually meant is that you could get even better at maths quicker. Because you could bypass the easy stuff at the bottom end and do more exciting things and more creative things with mathematics. And I think it's exactly the same as I say to my content team. Instead of writing the basic articles, they can write the opinion pieces, they can add their creativity, a level and only so many hours a week that people work. So if they can short circuit some of the more basic creative things that they have to waste their time doing. And get onto the more exciting creative talent, and that's the balance that I think that we are trying to look at to say, " Actually we want to embrace even more creativity and AI should give us more time to be more creative."

Megan Kresinske: Agreed. Abran, any thoughts?

Abran: There's so many barriers, there were a lot of barriers to entry to a lot of these creative suites. I know a lot of people who wanted to be graphic designers, they couldn't afford the Adobe suite and a lot of things. And there are a lot of people who have the imagination to be like a world builder, like a James Cameron or a George Lucas, but they don't have a studio backing them. But these tools like RunwayML and Midjourney and ChatGPT give you the tools to be like a world builder and create your own indie projects or spec work. Or even ads and just do something on your own with your own imagination. Where people get worried is that it was supposed to be given to the independent to have a tool to build that world. Not for the George Lucas' of the world to use these tools for shortcuts. And that's where the SAG- AFTRA conflict is happening is, " It wasn't meant for you to replace us, it was meant for little folks to get their shot."

Megan Kresinske: A hundred percent. Yeah, I listened to a podcast the other day where someone said, " People looking to use ChatGPT for the exact answers are obviously using it wrong, right? We should be using it to come up with things we may not have thought of and then bring in the humans to complete those ideas to full adoption." So I have hope, Gareth, that we're not going to live in a sea of the same content if humans are still involved. Yeah.

Gareth: I think it short-circuited, I helped my daughter write her resume for university that she's done over this summer. And what we did is she wrote one herself and then we put the right prompts into ChatGPT to write. And actually what it did is it helped her think about some of the ways the sentences were constructed, some of the ways that she wanted to put it across. It didn't write the perfect resume to get into university, but what it did do is it challenged some of the ways that she was doing it-

Megan Kresinske: Your thinking.

Gareth: ...rather than an opinion from somebody else. It was just another opinion, but it was quite eloquent and well grammatically put together in her opinions which-

Megan Kresinske: That part is nice.

Abran: I'm a Puerto Rican, so I tell my team, " Throw your adobo on top of it afterwards. If you're going to use it to write your email, that's cool, but throw your adobo, throw your seasoning on top of it."

Megan Kresinske: Absolutely. Okay, let's talk about ethical AI usage. So as AI becomes more prominent, there's a lot of discussions happening regarding AI governance at large. So how can brands ensure they're using AI ethically and responsibly, especially when it comes to customer data privacy and personalization? Gary, can you kick us off?

Gary Pak: Yeah, that's a really hot topic. AI is only good in terms of what data you feed into it. And so if it's going out to the landscape and only looking at the things that the majority of the people are sourcing, then it's really going to shift towards the majority. And so I think there needs to be a lot of governance and moderation around evening that up. Where you input multiple kind of voices, multiple signals into your AI engine so that it has a much more holistic view and output on things. In addition to that, I think there's some, again, privacy issues around like, " Hey, what information are you actually giving to AI? Is it your personal information? Is it your financial information?" And just putting guardrails around that type of stuff so that it's not... again, it doesn't go off on a tangent and start revealing things that it shouldn't be doing and intruding in people's privacy and stuff. So I think there's some ethical governance moderation that needs to be implemented by every company to have oversight over how they implement AI and what the output of AI is.

Megan Kresinske: Yeah. Abran,

Abran: Another shameless plug. We just released a tool called DEIGPT for marketers and organizations to use to help scale the work of DEI within the company. But also as they put out assets, whether it's marketing materials or if they're creating a campaign to run it through this AI system that was trained by DEI experts and accessibility experts. To just ensure that you're not missing anyone. Or I hate to use the harsh term, but cancel proofing your campaign. Just check it through the system first, see what maybe you have missed in brainstorming this idea and this strategy and this campaign. And to have this be a collection of experts that would consult with you, but having all of them on demand basically through an AI tool. So the products that we put out we're very social impact focused. So if it's something that we're putting out aside from our client work, it's something that has to benefit positive use cases of AI. But clean data, so if you're using your own, just do your own homework on is it a clean data set? Read the white paper, where did they source their data from? Sometimes you fall for the sexy UI of a new platform that's really just like an open source tool made by a...

Megan Kresinske: 20 year old inaudible.

Abran: Yeah, basically. And they're just like web scraping without really taking that into consideration of where the data's coming from.

Megan Kresinske: Well, especially because AI is only as powerful as the data behind it. So if these are new companies, it's like how powerful can they actually be?

Gary Pak: There's a company I won't mention, but they produce a lot of content in the financial industry, so mortgage rates and things like that. And they completely leveraged AI to auto- generated content kind of unfettered and unmoderated. What ended up happening is that the AI generated false or inaccurate mortgage rates and displayed to consumers. And now there's an FTC lawsuit against this company in the range of like$ 500 million that they're looking to seek. So again, I implore everyone in this room-

Abran: Gives me heart inaudible.

Gary Pak: Yeah. Moderation and governance around what your AI outputs is really key, because again, if it goes unfettered, it could lead to inaccuracies that violate certain regulatory and governance issues.

Abran: So here's why they did that, there's this new concept called Auto Agents, right? Creating automated set it and forget it use cases, that these 20 year olds came up with. They're like, " Oh, you can automate ChatGPT and just let it do its own thing." But then it goes left, there's things called prompt drift and it hallucinates, and you always have to keep that human in the loop to put that final touch. Even if it's creating the bulk of the content, you always have to have a final human editor to make sure that all that is accurate.

Megan Kresinske: And even though that was a scary example, I think it's a good example and actually leads to my next question that I'd like to hear from you all, and I'm sure the audience will as well. How do you all stay up to date with everything going on in the world of AI? Any recommendations to the audience? I know I dabble in podcasts, but how are you all the experts here? Abran, maybe I'll start with you since you've been doing this the longest.

Abran: So my secret is I hire a lot of 19 and 20 year olds that have a lot of time to... and they enjoy reading the white papers, demoing, doing the trial and error, giving you proof of concepts of stuff, and then figuring out how does this apply to my industry. There were folks that started with me in 19, and now they're working at OpenAI at 21, 22. So the younger, the better. And Future Tools is a great resource. Matt Wolf is great if you watch his YouTube videos or go to his website, he does a great job of curating what the latest news is. It's something that I reference a lot as well.

Megan Kresinske: Very cool. Gareth.

Gareth: There's so much. I'm in a conversation in the bar I think last night. I think you've got to chase down rabbit holes sometimes with the curiosity. You read an article that leads to another, there's a citation to another article, there's a link to something else. I find some of the best bits of information are when I start off reading one article, but then I'm down a rabbit hole over here on something else. And it gives you having a curious approach to it, I guess, and open- minded to it has always been it. One of my go- to's is always Amy Webb's stuff. I think she does a good entry point for a lot of people. She does the Future Institute piece of work. If you've not come across that, it's really worth getting, is a free resource around technology. It covers all verticals. She's a big speaker in Austin here at South by... Every year she's like a poster child of South by Southwest now. But she's grown that from, there was 20 people in a room three years ago when she did that report. And I think it's a good starting point for these things where you can pick your vertical and what's relevant for you in AI, and it's a launchpad from it.

Gary Pak: I've been fortunate because we have a network in private equity in VC, and so we understand emerging companies that recently got funded, so I would encourage everyone. A lot of this is publicly available to see which companies are being funded and the synopsis of how they're applying AI. But I would also recommend use your network, your vendor network. AWS for example, I think a lot of folks in this room probably partnered with AWS in some sort of capacity. They have insights into a bevy of applications of AI across all the startups and companies that are leveraging AI to drive the business. So use your network for sure.

Megan Kresinske: Nice. Well, we've officially restarted the clock, so we need to get off the stage. But I want to thank Gary, Gareth and Abran for joining me. If you want to talk AI, come find us. Thanks everyone.