Lightly edited for readability. Niall, you have j- uh, recently launched your new venture, The Creative Engineers.
Speaker 3: That's right.
Speaker 2: What is it, and what's the problem you solve?
Speaker 3: So, The Creative Engineers is a new consultancy. I founded it with three other advertising and marketing veterans. Uh, we launched last week, so great time to be speaking to us. And the problem we're trying to solve is that marketers need to produce ever more content. I was at an event yesterday, and the CMO of Vanguard said, "You're always marketing more. You're never marketing less." And so, a key challenge for marketers is getting enough content to reach the right people in the right place at the right time with the right degree of personalization. But marketing budgets are not going up, so the question is, how do you create that content? How do you generate the relevance? And our answer is you need to have a better content operating model. So, how do you connect the people, the processes, the tools to be able to produce more? And obviously, AI is a massive catalyst and opportunity to be able to do that better, uh, and increase the, the volume and the quality of what you do.
Speaker 2: And how do you actually do that?
Speaker 3: Great question. So, we are already doing it for a few people. So, last week there was an announcement that we're working with Olwen, the National Lottery. Uh, and so we look at it in three phases. So, one is an initial audit, uh, and, uh, then a scoping. So, identifying, uh, what are the use cases, what are the workflows that you want to optimize? What are the outputs you're trying to get? What's the right volume of content you're trying to produce? What are the formats of that content? What tools do you already have? What people do you already have? So, what's your current state? Then we figure out, to get to the levels of content creation and production that you're looking for, what's your future state, and what, what are the right workflows, for example, to have in-house? What are the right workflows to have with an agency? What are the right tools to create those workflows, and what are the skillsets you need to execute those workflows as well? So, that's kind of the easy bit. The hard bit is going from one to another, uh, especially if you're being ambitious and you're trying to change the way you operate in quite a big way. And I think that's where we do something quite different, uh, which is we don't give people some slides, um, and then say bye-bye. Um, we sit in their office with them. We're helping them write the job descriptions. We're training them on the tools that they're using. We're helping them put guardrails in place on their AI use cases, and we're making sure everything works. So, that means that it could take three, six, 12 months to actually implement the program, but the key thing is we're making sure that everything works and that business as usual can still be successful even though you're changing the way you operate in quite a big way.
Speaker 2: And you've obviously been working in this sector for, um, a while.
Speaker 3: We, we, we both know it's been a while, Justin.
Speaker 2: Thank you for not mentioning how long. Um, why don't you just kind of run me through sort of, I suppose, the journey to date, and kind of those key milestones and moments that have really helped define you and your career?
Speaker 3: Yeah. Thanks. Uh, I think, uh, the, the... this has been a kind of a long time coming. So, uh, I've been a, a marketer, uh, for, for a long time, starting off at places like P&G, Time Warner, Travelocity, and then running a digital transformation agency called The Knowledge Engineers. Um, and so, you know, we've seen the opportunity on the media side to reach your target audience in a much more personalized or segmented way. The kind of the media opportunity to be much more granular in your targeting has always been there. And as I've been helping, you know, large marketing teams take advantage of that, one of the key blockers has always been the content. So, can you really take advantage of the opportunity that digital media gives you when you don't have enough content, or that content isn't purposed in the right way for the granularity that you can get through the targeting that media offers? So, I then, uh, sold that business, uh, in 2017. Uh, worked, uh, for the buyer called Evardo. And then when I left there, ended up, uh, accidentally working in AI, uh, before it was cool, for a big e-commerce, uh, e-com AI startup called Black Crow. Um, so raised money in 2021 and again in 2022, um, to, uh, actually capture intent using machine learning. And then, again, that problem arose of how do you give... how do you use the signals that you've captured in a really granular way to give people a very personalized experience when you can't, uh, give them the exact content that, that the signals would e-expect you to? And so, when generative AI arrived, here lay the solution to the problem that I'd been encountering for, for really the last 10 or 15 years. And then I met, uh, Helen, Chris, uh, and Morgan, who are my partners in The Creative Engineers. They've been helping organizations, uh, with these types of problems for a long time, and really what we brought together is kind of classical, uh, understanding of production methods along with my understanding of AI. Put those two things together to say companies can now create all the content that they need to be as efficient as they can be, whilst at the same time obviously protecting the quality of the creative work. It's absolutely not about mass production to the, um, to the sort of be- lack of benefit to the quality. Uh, and it's also about making sure that you're still on brand, so you're still building the brand even though you're producing more content. Uh, and that, and that's the tricky bit. I guess that's where my background comes together. I've always been very proud brand marketer, both B2C and B2B. Uh, and it is so important as you're using AI to extend the quantity of content to make sure you're still producing- Quality of content as well and, and you're still creatively proud of everything that comes out.
Speaker 2: And, you know, as, as you've sort of navigated through various kind of paradigm shifts, whatever you wanna kinda call them, right, in terms of how this, the landscape has evolved over your, your career, what's particularly kind of different about this time?
Speaker 3: Yeah. Great, great question. So I think, um, obviously speed is, is different. Uh, and I'd actually say, uh, I've been working and training marketers ever since gen AI kind of happened and ChatGPT came out. Uh, and I would say the last four months have been unbelievably accelerated in the impact of gen AI on marketing teams. So before it was, uh, hard to build really good use cases, it was quite time-consuming. You really had to know a lot about prompts. Even prompting something like nana banana for images took, you know, real experience and knowledge of, of prompting and, and prompting for that purpose. But what we've seen through, uh, Claude, uh, Claude Cowork, other orchestration platforms like Writer.com, is the ability for marketers to create agentic workflows and automations kind of out of the box without really, really specialist knowledge, without having to be an expert prompter, that can actually deliver quite complex workflows, and I think that that's really only happened in the last, uh, four months. So I think, uh, speed, um, is, is really distinctive, and then clearly the kind of the nature of where you can use gen AI, whereas, um, with the growth of digital and machine learning, it was very much focused on the media side. It didn't really touch creative that much. Um, and now we can see that the impact on creative in a generative AI world is very, very big, and that the number of modalities that's impacting is, is growing. So starting out with text, moving on to pictures last year, and I think this year we'll, we'll see kind of the acceptance that AI-generated video, even 30-second video, can be a really high-quality format if you know what you're doing and you're prompting in the right way and you're using the, the right tools. So I think it's, it's moving more quickly, and it's impacting different places in marketing than kind of previous transformations did.
Speaker 2: And you talk about impact there. Give me sort of a sense of kind of what impact you're seeing as a result of the work that you're doing.
Speaker 3: Yeah. So I think when you look at, um, couple of things, uh, think you can probably think about, uh, operational efficiency is one area of impact, and you can think about creative output and the volume and velocity of creative output as the other. So when you think about operational efficiency, there are definitely use cases where you can really significantly automate some of the most thankless tasks in marketing. So a really good example I've been working on with one of our clients is brand guardianship. And so when you think about a large organization with many different people creating content, creating emails, creating internal comms, all of that needs to go through a brand guardianship team. Um, it's not the, you know, the best part of a brand team is doing that brand guardianship job, and many of the inputs into that process are, you know, relatively simple to say, "No, that's not quite right." So the wrong font, the wrong color, the wrong use of the logo, the wrong use of spacing. So if you can automate some of that, not all, you can't automate all of it, but if you can automate 50% or 70% of that and give immediate feedback to the submitter to say, "You've maybe made a bit of a mistake here. Go back to the brand guidelines. You need to look at this bit," you can significantly take some of the load off that brand team, and then they can invest more time in actually doing the work they want to, which is creating great, great brand marketing. So I think on the operational process side, that, that's a great opportunity to, to save time, to save money, and, and actually improve the quality of people's jobs. On the velocity or, and volume of content, they're I think relatively early days. Uh, we're working with a, a great startup, um, called Leapfrog, uh, who have got a really good approach to, uh, photo generation, synthetic talent that enable some of our clients to do, um, adaption and versioning much more quickly than they would've done otherwise. They've got, uh, a great client list. Not sure how much I can say about who they work with, but they've got very well-known, uh, household name fashion brands where l- you know, large parts of their digital marketing is populated with, uh, synthetic talent or talent which has got a degree of, um, synthetic included in it. So what does that mean? It means you can create a lot more marketing assets that are a lot more relevant for the audience. So perhaps you're entering a new market, you're a fashion brand that's entering Thailand, and, and let's say you don't have time to do, uh, a shoot for Thailand. You don't have time to create all of the different assets for Thailand, but you can use your virtual talent or your synthetic talent or synthetic shoots to rapidly version, uh, a, a 2,000 product description pages for that market with all of your products, um, looking as if they're created for a local market. So enables you to enter new markets more quickly, enter, create, enables you to create engagement with people that otherwise you wouldn't get good engagement with. Um, so I, I, I think that's a, a massive opportunity, but always important to say with the right creative guardrails in place, the right brand guardrails in place, and, and make sure that everything's still on brand and high quality.
Speaker 2: And to your point around the pace, you know, of, of change in that four months you sort of just described, and you mentioned, uh, Leapfrog, which I'm gonna have to check out now, thank you very much, and I'm sure everyone else listening is gonna be very grateful, too. Um, how do you also sort of, you know, um, go through that process of identifying, um, the tools, right? Because, you know, in a world of, you know, constant evolution, you know, the re- the level of kind of potential redundancy of those tools and the leapfrogging of Leapfrog, I suppose, in a sense, you know, is, is, is a, is a, is a risk, right?
Speaker 3: Yes. I- I- It's a, it's a real challenge. Uh, just before this podcast, I was at, in, in a meeting with the AI team of one of our clients exactly talking about how, uh, the tools that they thought they might be using today are now different, um, than they, than they thought. So I think you've got to, um, you can't drive yourself crazy so you've gotta be a bit sensible about how you do that. Uh, for sure, you know, if before you might have signed three-year deals with vendors, you definitely shouldn't be signing three-year deals with vendors. You know, one-year rolling contracts is probably the maximum you should be thinking about, so give yourself flexibility so if a new tool comes along that is better than what you're using, you, you can actually switch without, uh, significant financial penalties. Keep, keep your eyes open. There are some great communities in, uh, both AI and in marketing and in places like X where you can just keep yourself up to speed, um, with tools that seem to have traction. Always be looking for people who can demonstrate real use cases. A- and there is a little bit of a lag when we think about kind of corporate use. Whether you're B2C or B2B, you still wanna make sure that the tools you use are, you know, SOX compliant or compliant with your IT policies. You gotta make sure that they're robust from a legal and copyright perspective. So it's not like, uh, most marketers are gonna pick up something that's sprung from a vibe coding project yesterday. They, they're gonna have a little bit more of a run-up, um, and so you- you're gonna get a bit more visibility of other people using that tool before, before you do. But, I mean, I think just, you know, the biggest advice on, on identifying those tools and figuring out what works best is you, you have to personally engage with those tools. You have to use them. Uh, you have to try them out, and if you have to have a team of people if you're working with a team, um, who are also inspired to do that experimentation, um, otherwise you, you know, you really can get left behind very quickly or, or not understand where you should be investing.
Speaker 2: Now, you're someone who really understands the kinda human side of this, right? I remember a conversation we were having when you were, um, bringing in fantastic kinda capability building and diagnostics on k- existing kinda capabilities for parts of the kind of the, the WPP sort of, uh, network.
Speaker 3: Yep.
Speaker 2: Um, and then you were taking those folks through, you know, a really transformational kinda journey, uh, um, as you went through those businesses. Tell me a little bit about sort of like how you're approaching that now, uh, you know, on the human side.
Speaker 3: Yeah, great, great, uh, point. Uh, so I think within the Creative Engineers business, um, we're very focused on the use cases, so what are the skills you need to execute those use cases, uh, and are you getting the most out of your tools? So one of the biggest issues we see when we go into new clients is they might have very expensive software tools, and they might think that they need to buy more. But it, it might actually be they just need to use the tools they've got better. And so I think what we, we target our human capability on very specific actions. It's not generally big groups of people. It's very focused on particular, uh, workflows. The one exception to that is I think you need to have a leadership team, like whether it's a marketing leadership team or a company leadership team, who knows enough about AI, the opportunities of AI, where AI is now, in order to be able to give you the support that you need to have a strategy as well that you're feeding up into. Uh, but I do have another hat, um, so I do chair an AI education company called Edify, um, and Edify is actually working with quite a lot of, um, agencies, uh, again. Um, a- and that's a, a bit different, um, so they are focused on much broader sweeps of, of kind of human skill building. And I think probably the most interesting area that, that they're working on at the moment is really focusing on managers. So if you focus on managers, uh, they are the people who really understand the workflows. They understand how the work gets done now. They understand the tasks in those workflows and the tools that are used for those tasks. So if you can give managers a degree of competence in AI, and you can give them a bit of understanding of how do you identify a use case, how do you decompose a workflow, 'cause it's, it's not really that hard, then they can start to assess their use cases. They can brainstorm with their team. They can support people who are doing experiments because they can understand whether that experiment's valuable or not. So I think, um, managers play a really important role in kind of medium and large organizations in catalyzing the AI opportunity.
Speaker 2: And taking all that for granted, which I think, you know, is a, is a great answer to my question, but below that, there are obviously some fundamental disciplines, capabilities, whatever you wanna c- people call them competencies. People call-
Speaker 3: Mm-hmm
Speaker 2: ... whatever, whatever you want to call them, and I, I have no... I will not judge you, I promise. Um, what, what are some of those kinda more fundamentals that you feel like kind of this next generation of the workforce needs to have?
Speaker 3: Yeah, it's a really good question. I mean, uh, it's a question I ask myself. I've got a 10-year-old and a 12-year-old, and so thinking about, you know, what, what skills are they gonna n- need in the future? Obviously still a bit early to think about jobs, so you gotta think about what, what are the, what are the skills and mentalities you want to build in those people. So I, I think it's a really hard question to answer from a skill set perspective because clearly something like, you know, computer science or coding is- are not kind of vertical disciplines that you necessarily need to be an expert in anymore. My family have built an, an app, for example, that w- uh, it's in the queue to be reviewed for the app store, a family, a family quiz, uh, app. None of us can code. None of us can design . And you'll probably tell that from when, when it, when it gets into the store. Um, but I think, uh, critical thinking is a- absolutely essential, so being switched on, reviewing what comes back from an LLM, critically assessing that, understanding how to iterate, so not just accepting what comes back the first time but knowing that you have to iterate. Uh, persistence, so knowing that some of the things you're doing are gonna fail. Sometimes the LLM's gonna get it wrong, so you gotta be persistent. I think if you think back to maybe 10 years ago, there was like a big trend around the b- the growth mindset. And so maybe it's not so much around, um, skills. Maybe it's more around a, a mental approach, and that you've gotta try lots of things, you've gotta put a bit of effort in.
Speaker 2: Mm.
Speaker 3: Some of those things are gonna go wrong, and you're constantly learning about what works and what doesn't. We work with, um, some AI engineers that we, in the trendy way of saying it, forward deploy into our, our clients, and none of them have a traditional computer science background, but they are amongst the most accomplished users of ChatGPT or Claude Code. They have just worked really hard at using it, applying it to their own use cases, learning from all of that, and constantly keeping those skills, um, up to speed. So I think you, you can definitely have some kind of broady human, uh, skills, uh, but it may be it's more of a mindset than a, than a skill set.
Speaker 3: Um, I mean, I think if, if you, if you look at what Anthropic have delivered in the last four months, I saw a, a calendar visualization of it, um, you, y- you've gotta be unbelievably inspired by the progress that they've made in that product, particularly as business users of that product. The series of features that they launched from the ability to access your machine and get it to carry out tasks from your phone, from launching CoWork, so giving kind of non-coders a terminal that they can work with and produce, uh, content, websites, and lots of different marketing-type, uh, tasks, uh, to computer use, which they launched a, a couple of days ago. So I think that, that kind of shows you two things. One , one, a group of people who are working very, very hard, but to a very clear strategy, and secondly, they're using AI to help them build some of those things. And if you go through some of their tweets, they're actually now very transparent in the systems and processes they set up, um, to help them, um, create good work, but using AI to create that, that good work. They kind of... They're proof of their own, um-
Speaker 2: They're practicing what they preach.
Speaker 3: Yeah, they are definitely. I was gonna say they, they eat their own dog food, but I think you put it much more artfully. Um, yeah, so I think, I think, I think that's really impressive. I think from a marketing and advertising perspective, a lot of people are still in, um, early stages. I think there's some good, uh, use cases. So yesterday I was talking here, I think Unilever have done some really good things around digital twins and changing the way they create production in their, uh, hair care team. They've used, um, AI assistants for briefing. It reduced the briefing time, improved the quality of briefing. Uh, I think, uh, TransUnion have done some really interesting work. I think, um, anyone who's been brave and experimented with AI video, like, um, Calci, uh, last year at the NBA finals. They arguably not the most perfect, you know, ad creative in the world, but they, they gave it a go and did something very, very quickly. So I'm not sure I can see anyone who's sort of put together a sustained end-to-end use of AI that you would go, "Oh, that's the perfect kind of toolkit." But I think people are beginning to string together use cases in a consistent way that, that are beginning to be, um, im- impressive.
Speaker 2: Yeah, I was talking to someone this week about, um, Lovable. Again, a- another sort of, you know, um, AI business which is, as y- I'll use yours, eating its own dog food. Um, and, you know, they are, you know, uh, redefining or recalibrating their proposition on a quarterly basis, right?
Speaker 3: Yes.
Speaker 2: 'Cause back to your point around speed.
Speaker 3: Yeah.
Speaker 2: Um, and when you are sort of, you know, listening to this conversation, um, and thinking, "My gosh, you know, this is great, but, um, where do I start?" What's the answer?
Speaker 3: Yeah. I read a good, great tweet the other day that said, "AI seems to be moving so quickly at the moment, the only way to keep up is to be unemployed." So I think the, the first thing to do is to say it's gonna take... You know, you need to allocate a bit of, a bit of time. I think by far the, the, the quicket- the quickest way today as an individual, if you're not tied into a corporate IT policy, and you want to experiment on your own, I'm not encouraging anyone to use shadow AI at work, so I'll, I'll get told off by our clients. Um, but, uh, if I was doing it as an individual, I, I would get Claude, I'd get Claude Code, I'd get Claude CoWork. I'd go and look at what's the best practice to get set up. I'd think about some of my personal use cases, and I'd build some skills in Claude for those use cases. Then I'd start turning those skills into agents, and then I'd start learning how to build and control, um, agents. And you can probably do that in three hours. And, and once you start to build and evaluate agents, like, you've got a, you've got a long you've got a long way. Um, so I, I, I would just get started with, with hands-on usage, and it's not hard to track down some of the best practice ways to get started.
Speaker 2: And I guess the flip side of that is I am working inside a, a, a larger corporate, uh, enterprise brand, and, um, I still wanna do the same thing, and maybe I've gotta convince my manager, who hasn't yet had the benefit of working with you, or maybe I'm the, the CMO, and I've gotta convince my board, who are sort of asking me what I'm doing in AI. But, you know, I, you know, I basically need to present some form of business investment case. What do I do in those sorts of situations?
Speaker 3: Yeah, great, um, great question. So I think, um, sh- show and tell is definitely, um, the way forward. I mean, I, I think for most managers, especially if you go to the C-suite, uh, I think most CEOs and people in the C-suite are very aware of how fast the world is changing and what that might mean to them. I think probably it's not so much that they're not interested in AI opportunities. Um, it might... The problem actually might be more that they are looking at it through the wrong, the wrong lens. So maybe their 100% focus is, "We can use AI to be more efficient and cut costs." And you can use AI to be more efficient and cut costs, but is that the best use of AI in your business? Is that the best way to build your business? So maybe, uh, I, I think if you're, if that person and you're talking to your manager or to your, your C-suite, it's about really thinking through where can we get value from AI? What tools do we need to get that value? To show them a couple of things that you've done, um, without using, say, confidential information if you're in a, a big company, a- and, and prove to them, um, what, what can be done. Um, but I think, uh, i- in my experience, it's not, it's n- not a hard push at the moment because they're all reading the FT, they're all reading The Wall Street Journal, and they're all very much aware that the world is changing. If you work for a listed company, in every analyst call, the analysts are asking you, "What's your AI strategy? What are you doing with AI?" Um, so I, I don't think it's a, a difficult s- pitch. I think getting the direction right is maybe a bit more difficult.
Speaker 2: Last, last sort of point now really, uh, before a couple of quick-fire questions, so, um, get ready. Um, creativity. I was talking to CMO yesterday, and she said, "We issued a brief, and basically we got... You know, we shortlisted four, and then we basically ran the brief through, uh, an AI product, and basically every single one of the responses basically mirrored the, the, the output." And you're laughing, um, and, um, I wanna kind of ask you, well, first of all, why are you laughing? But I think, I think I can understand, right? Uh, but more importantly, um, you know, how are we gonna kind of move on from, from this? And how, how do we rediscover kind of AI-powered, AI-fueled, AI-augmentation of, of creativity pushing us to new heights?
Speaker 3: Yeah. Look, it's a, it's a... I'm laughing because it's such a, uh, ob- it's such an obvious thing to, to do, uh, and it's so obvious that you're probably gonna get caught. It's like copying and pasting some writing-
Speaker 2: Homework
Speaker 3: ... that's come from ChatGPT, and it's, it's got some em dashes and some arrows in, and you're like, "You just copied and pasted that from Chat- ChatGPT." I, I spoke, uh, in another example, uh, spoke to a client in a listed company, uh, and someone had brought up examples of their environmental, um, uh, policy implementation, and they'd just put it into ChatGPT, copied and pasted it, and it'd gone, gone into the draft of the annual report, and some of it was complete hallucination. And they said, "Where," like, "Where did this come from?" And they tried to pretend that it was real, and they're like, "Yeah, I just used Ch- ChatGPT." So, so I think, number one, that shows a very sort of primitive approach to using, uh, AI. I think in the process of creativity, we're not looking for AI to solve the problems for us. We're looking for AI to help us explore the problems, help us explore approaches to the problems, uh, help us get started sometimes when we're thinking about creative solutions to them. But no, if you, if you break the process of creativity sort of down into the workflow, the simple way of doing it is to say, "I, I, you know, I need five creative ideas to meet this brief. Go." Those ideas will not be very good. If you... But if you break it down and say, "First of all, I want my planning agent to really help me understand the consumer. Here's some data I've got. You go and find some other data. What are some different insights?" Or, or take a, a planning methodology that you like, that you used before, and you use that planning methodology to help you come up with insights, but you accelerate it by using AI. Then you take those insights and say, "Let's be, you know, be a creative partner. I've got these ideas. I think this is interesting. What are some other interesting things? What would, what, what have I come up with that, um, that could be built on? Or what have I not come up with that, that might be obvious?" And then you can take that through the whole creative development, uh, process. So I think, uh, if you're using it in that assistive way, if you're using it with your frameworks, your approach to creativity, and then you're imposing your judgment and your taste, and, and, and what comes back will very rarely at the highest level be, be a great answer. But what it can give you is a starting point that you as a human can turn into a great answer to a, a creative brief or a strategic, uh, brief. So I think, you know, there, there's using AI, and then there's just being bloody lazy. And I think what you just described is probably, uh, the s- the second.
Speaker 2: Okay, wrapping up now. Um, what's the one piece of AI that you couldn't live without?
Speaker 3: Right now, for me, it's CoWork, Claude CoWork. It is, it is running 24/7. I, I've hated Macs since I used them as a student, uh, in the 1990s, and I've always sworn never to go to a Mac, and I'm gonna have to go to a Mac because my PC cannot keep up with Claude. Um, so it, it... I, I'm, I'm ashamed. I'm, I'm s- I'm annoyed with myself. I feel sick. But next week, I'm gonna be buying a Mac.
Speaker 2: Well, as an Apple fanboy, all I can say is welcome to the family.
Speaker 3: Thank you. Thank you. I feel welcome.
Speaker 2: And you mentioned, uh, a couple of, um, sort of sources of inspiration. Could you be so kind as to kind of share what are your kind of go-to kind of resources to keep abreast of who, who and what and how this world is evolving?
Speaker 3: Yeah, I mean, I'm not necessarily, um, the biggest, uh, fanboy, but I think X, um, for AI is, is pretty good, uh, in terms of people talking about what feature's working, what's not working. I think some of the frontier models, uh, like, uh, Google, Anthropic, and, uh, ChatGPT are now being very communicative on X, so you can learn quite a lot quite quickly. There's quite a lot of deep dives coming from not only those people but really, um, you know, very serious people in data science who've been around this world for a long time, like Andrej Karpathy. Um, so I think A- A- X is probably a really good go-to place. There's some really good communities on Reddit. Uh, and then I think, you know, there are some private communities, uh, run by some of the, uh, tech, um, vendors, some of the tech companies. Some, um, industry organizations do have communities, so I think a Slack community or, or Teams community is another, a great, great place to learn.
Speaker 2: Niall, you've been so generous, and you've also been incredibly humble, um, answering my questions, and I'm really, really grateful. Thank you for an amazing conversation.
Speaker 3: My pleasure. Thanks, Justin.