Latest / Inspiring Futures - Lessons from the Worlds of Marketing and Advertising / Lior Eldan- Co-Founder- Moburst- The Agency of the Future
Transcript
- 0:00SPEAKER_00: Welcome to the latest episode of Inspiring Futures.
- 0:03SPEAKER_00: My guest today, Leo Eldon of Moburst.
- 0:05SPEAKER_00: Welcome to the podcast.
- 0:08SPEAKER_02: Thanks, Ed.
- 0:08SPEAKER_02: It's great to be here.
- 0:11SPEAKER_00: Perhaps you could give us a little brief introduction into your sort of career trajectory and then a little bit about the company.
- 0:20SPEAKER_02: Sure.
- 0:22SPEAKER_02: So I uh started off uh um in the early days of SEO when Google uh was this wild west and uh everyone was kind of figuring out how to uh influence the ranking, so started off in SEO.
- 0:38SPEAKER_02: Um uh from there I went into um online reputation management.
- 0:44SPEAKER_02: So rather than taking a website and promoting it, taking a brand or a person and owning the search results.
- 0:51SPEAKER_02: Um and after that, I uh uh I knew my next uh my next move is into mobile.
- 0:58SPEAKER_02: Uh mobile was booming, it was 2013.
- 1:00SPEAKER_02: I I hooked up with uh a friend uh from the army, um and uh we uh founded uh Moberist, um, which we started off as a mobile agency.
- 1:13SPEAKER_02: Um and 13 years later, today we're uh uh we did uh five acquisitions today.
- 1:19SPEAKER_02: It started off at mobile, grew our presence to work with a bunch of um um biggest brands in the world essentially.
- 1:27SPEAKER_02: So we uh dealt with all of Google's B2C apps and uh worked with uh Dropbox and Microsoft and Samsung, etc.
- 1:36SPEAKER_02: Um, and in the last uh six years we we expanded uh to do much more than mobile.
- 1:42SPEAKER_02: Today we're uh a full service uh digital agency, still uh uh very skilled and prof and mobile first, uh, but deal with anything from PR to influencer marketing to uh BI services to Amazon marketing, we're 150 people globally um with a wide wide range of uh clients in all verticals and all sizes.
- 2:06SPEAKER_00: So um take us through the sort of uh inflection point when you decided that you need to do expand beyond the mobile frame.
- 2:19SPEAKER_02: Yeah, um so it was around 20 uh 18, 2019, we um established a pretty decent presence in terms of uh mobile.
- 2:30SPEAKER_02: We were kind of getting invited uh by all these big brands to their RFPs.
- 2:35SPEAKER_02: Um, but we we understood that the amount we were very kind of premium, only working with top-tier, you know, fortune, whatever, 1000, 2000 companies.
- 2:45SPEAKER_02: Uh, and we figured that the amount of such companies uh which are doing uh primarily mobile and looking for a mobile agency in the US is quite limiting.
- 2:56SPEAKER_02: Uh and we also had uh clients kind of asking us if we can help them with their website, if we can help them with other challenges across the digital marketing.
- 3:05SPEAKER_02: Uh so that's when we we kind of uh uh thought, okay, let's let's try and expand uh beyond.
- 3:11SPEAKER_02: Um and uh there were certain uh things which were you know strategic.
- 3:16SPEAKER_02: We knew we wanted to do PR because we had clients and prospects asking us uh consistently, but there were things which we kind of stumbled on once we started to do MA more seriously and had uh uh we have a full-time corp dev guy nowadays.
- 3:32SPEAKER_02: Uh so there are kind of services which we stumbled on, good companies that we thought there's there's uh uh it makes sense to add those uh capabilities.
- 3:41SPEAKER_02: So podcast booking and uh and a few other stuff which were more uh um uh opportunistic by by nature.
- 3:49SPEAKER_02: Um but uh uh but yeah, we we did that, and also we we uh there are several services we uh we decided to build in-house, like influencer marketing and others.
- 4:00SPEAKER_02: Uh so it's uh really depending on each service if we find a good company to acquire and what we think uh we need to serve our clients better.
- 4:09SPEAKER_00: Yeah, makes sense.
- 4:11SPEAKER_00: So if you if you go back to 20, if you could contrast very as simply as possible the transformation in marketing from 2013 to 2026.
- 4:20SPEAKER_00: So kind of what what have we gone from and where are we at now?
- 4:24SPEAKER_00: Because that seems to be very um much the driver of why you've changed, why you've evolved, and kind of where you started.
- 4:33SPEAKER_02: Yeah, it's it's interesting because in uh when so we we when we started off, uh App Stores was a new phenomenon and kind of how to get apps uh to appear, to rank, to get users to actually download apps was back in the days when people would actually look for new apps and cool stuff to download.
- 4:53SPEAKER_02: Um uh and and and we essentially are are one of those that really established um ASO app store optimization as a practice.
- 5:02SPEAKER_02: Um and uh um and it was very exciting because there wasn't any cookbook, it was uh everything was very new.
- 5:12SPEAKER_02: Uh uh, so we we kind of felt we were you know building the practice uh and and uh the best practices by trial and error.
- 5:21SPEAKER_02: Um and fast forward to for today, I feel that uh when thinking about um what's going on with with uh with search and AI search and uh what's being now uh uh coined as AEO, um uh which is answer engine optimization or geo uh generative engine optimization, um it's it's essentially I I see a lot of similarities just on a much wider scale.
- 5:49SPEAKER_02: So rather than you know just an app store, now the LLM's uh NAI looks at a brand's digital footprint at large.
- 5:57SPEAKER_02: So not only your website, um what do what do other uh websites say about you, third-party PR, uh what kind of podcasts are you and what videos speak about you?
- 6:08SPEAKER_02: Uh what do review sites, uh is there any kind of industry publications, what do folks on Reddit say about you?
- 6:14SPEAKER_02: So all these kind of different pieces of a brand's uh digital footprint.
- 6:19SPEAKER_02: And uh and one of the things we realized as we were kind of entering the space in the last uh uh two years is that we were actually we touch all of these digital uh uh all these different touch points.
- 6:33SPEAKER_02: Uh so we we really kind of uh uh see this as uh a new playing field, so to speak, in the market, uh where uh AEO is a relatively new thing.
- 6:45SPEAKER_02: Most brands don't even know what it is, or kind of know that if they ask ChatGPT uh or perplexity or their favorite LLM uh if they're there or not, but don't really know how to influence it.
- 6:58SPEAKER_02: So uh we we feel this is kind of uh in in our personal uh uh story, it's kind of like 2013, all over again on a much wider scale.
- 7:08SPEAKER_00: Yeah, it is because I mean if you think about if you think about what you were doing, it was a very specific thing back then.
- 7:15SPEAKER_00: And now, you know, in terms of like the breadth, you know, like the the it seemed quite quite narrow.
- 7:22SPEAKER_00: I mean yes, it was important, mobile was growing, everyone has billions of smartphones, and you know, app stores are key.
- 7:29SPEAKER_00: Um but now you're talking about things like the reputation of a company, um you know, and uh the skews that that company sells.
- 7:40SPEAKER_00: It sort of seems like a very much more broadened portfolio of uh problems that you're trying to solve for.
- 7:49SPEAKER_02: Yeah, because it's and one of the it's interesting that you're saying reputation, because it's uh uh it's not about you know being found that okay, now I'm looking for whatever uh company to do uh mobile marketing or or or for any type of service or product.
- 8:06SPEAKER_02: Uh and now I I get a short list of uh uh companies.
- 8:10SPEAKER_02: Now I start to ask.
- 8:11SPEAKER_02: That's the the usual kind of uh uh user journey, uh consumer journey to start to ask like, are these guys trustworthy?
- 8:17SPEAKER_02: How is their pricing?
- 8:19SPEAKER_02: Uh what are they good at?
- 8:20SPEAKER_02: Whether and what do other customers think about?
- 8:24SPEAKER_02: And that's where it's it's not just being seen, but actually how does the LLM, the AI, perceive you as a brand uh in terms of your the percept the perception, your reputation, as you mentioned.
- 8:35SPEAKER_02: So uh it's um I didn't even realize until we're speaking now, but this kind of ties to uh my earlier career around reputation management and and what uh an LLM, how it uh uh uh perceives uh a brand in terms of all the different uh aspects that uh might be interesting for consumers.
- 8:57SPEAKER_00: Yeah, and I mean I think this is kind of what we talked about earlier was this idea that how you've sort of got to where you've got to partly is um certainly with AI is is noticing the differences in the reputations of the companies that you've acquired and um kind of understanding how those work, which has given you intelligence that you can apply beyond those companies, right?
- 9:26SPEAKER_02: Yeah.
- 9:27SPEAKER_02: Um yeah, Henru, I think that that uh this is from a few angles.
- 9:34SPEAKER_02: One uh um you as you you you uh from the acquisitions we you know were exposed to other companies, to other founders, how they work, how they think, how we can kind of we can take some of their best practices, apply our own.
- 9:49SPEAKER_02: Um and and also from kind of uh uh marketing branding perspective, one of the things we we realized along the way is that we can't be kind of uh catch-all in terms of when uh you search from a good uh PR firm.
- 10:06SPEAKER_02: We can't be the results for PR and social media and media and marketing and all and mobile marketing and all of that.
- 10:12SPEAKER_02: Uh so we're we're kind of build we we we're taking more of a house of brands approach to to uh build each of the acquired uh entities as well as a few uh new brands we're launching to be kind of top of their game and can and control their niche rather than having one uh kind of uh catch-all brand, uh, which I think is much more difficult nowadays.
- 10:37SPEAKER_00: How do you how do you manage the hype from the reality?
- 10:42SPEAKER_00: I mean, you you you you pick up the marketing press, and you know, we talked earlier about everything's changing every day, and and you and you believe that there's a transformative revolution, but um things happen more slowly when you're talking to sort of average mainstream consumers.
- 10:57SPEAKER_00: So how do you how do you manage the delta between the kind of the hype and the reality?
- 11:02SPEAKER_02: Yeah.
- 11:03SPEAKER_02: Uh it's uh it's it's it's tough nowadays because if you yeah, if if you go into you know any any social media, LinkedIn or or X, uh there's definitely lots of hype and lots of FOMO.
- 11:18SPEAKER_02: Um I think uh we we definitely try and keep uh up to speed with everything that's going on on the one hand, uh but on the other hand, understand that we can't, it's we can't and we shouldn't really test every single you know new tool that comes out.
- 11:37SPEAKER_02: So we're we I think initially in kind of our AI um journey, there was a phase where we just tried and did everything.
- 11:45SPEAKER_02: Now we're you know, okay, so we have all the LLMs, but but we're we're kind of leaning more into uh into cloud in terms of how we work and the stuff we we build and the uh skills, um, and trying to find kind of okay where where do we go deep uh and where there might be okay, this is a tool where uh there there are capabilities for the creative team or uh for any speci any particular um team uh that a specific tool or mechanism or whatever is uh um so it's kind of be being open-minded enough to just to see if there's anything interesting uh out there, but also focused enough to kind of gain that depth and understanding and um and and try out.
- 12:30SPEAKER_02: So we're like I think the industry now making lots of uh uh um agents to help us with our daily work that can run analysis, that can kind of alert us on anomalies and stuff like that.
- 12:43SPEAKER_02: Um uh so there's you know the hype of just giving an agent to do everything, and then there's really the reality where now there's governance, what like how can we do it in a way which we're we're happy with the deliverable?
- 12:58SPEAKER_02: Is it client ready?
- 12:59SPEAKER_02: Does it uh uh hallucinate?
- 13:01SPEAKER_02: So uh that's kind of where I think we we there's that uh gap, I think, from how easy it seems on a demo or on a LinkedIn uh post rather than uh kind of day-to-day.
- 13:14SPEAKER_00: What do you what do you think the sort of given your exposure and given what you've been looking at as far as LLM is concerned and their role in the consumer journey, what do you think the brands have to do now that they weren't doing pre-LLM?
- 13:28SPEAKER_00: What's what have become the imperatives?
- 13:30SPEAKER_00: What have become the sort of critical things that if you if you continue to do what you used to do, you're not gonna you're not gonna be able to succeed in this world?
- 13:38SPEAKER_00: So how how would you identify kind of those things that that brands need to be doing?
- 13:45SPEAKER_02: Yeah, so for I think first is just add that to the measurement stack, right?
- 13:51SPEAKER_02: What is not measured cannot be managed.
- 13:53SPEAKER_02: Uh so uh uh um at the very simplistic end, just ask the the ask a few llms kind of on your brand or the the category.
- 14:04SPEAKER_02: I think in the more uh advanced, there are several tools.
- 14:08SPEAKER_02: We also built our our own uh uh tool that monitor the different LLMs uh across all these prompts to see what what brands are coming up, uh and kind of monitoring that on a daily or weekly basis to monitor changes.
- 14:23SPEAKER_00: I think that's kind of uh table stakes in terms of understanding the benchmarks intelligence layer, you know, just yeah, yeah, just having sort of benchmarks and intelligence layer, and then how do you go from there?
- 14:35SPEAKER_02: And then I think uh um and and then it's it's thinking more uh strategically on what you put out there uh in terms of uh and we're seeing this all the time, and I'm sure lots of uh your listeners as well, that uh I'm asking uh an LLM about my brand, and it's it's not really to the point.
- 14:56SPEAKER_02: I don't know where he actually got this detail, it might not, it might be what I was a few years ago where the product isn't.
- 15:02SPEAKER_02: Uh and and this happens because uh the LM uh score the internet takes all kinds of information, piece them together and and give uh the result.
- 15:11SPEAKER_02: And if uh it finds incoherent pieces that don't really match up, uh it essentially tries and takes the the average.
- 15:19SPEAKER_02: So I think one thing is to make sure that um to to make sure that whatever is communicated via social, PR, uh uh uh what you write on your own uh uh blog, etc.
- 15:32SPEAKER_02: That that's kind of aligned in terms of uh the messaging, in terms of the strategy.
- 15:38SPEAKER_02: Uh and and that relates to the our first uh my first my first point on the intelligence layer.
- 15:44SPEAKER_02: Like where am I strong?
- 15:46SPEAKER_02: Maybe where am I weak?
- 15:47SPEAKER_02: Maybe I want to hone the strategy of what I'm putting out there to start to speak about myself as an AI first company.
- 15:55SPEAKER_02: So I need to produce more content, which speaks about what I'm doing uh as it relates to AI.
- 16:00SPEAKER_02: So the LLMs would pick up that hey, I'm actually uh relevant for this uh uh for this type of query uh that the user asks.
- 16:08SPEAKER_02: Um and and then there are all kinds of there's all kinds of technical elements on stuff that you want on your website, you want schemas, and that's kind of the SEO part of AEO.
- 16:19SPEAKER_02: Uh so there are soon some nuances in how to get uh how to be more friendly and give them essentially the the structured data that you're they're looking for, FAQs, uh uh LLMs, TXT, although uh uh uh Google officially said it doesn't look at it, but but our testing shows that it does move the needle.
- 16:39SPEAKER_02: Um and and then there are other stuff which uh which have effect PR we've seen uh work works quite well in terms of if you now get uh uh an article or commentary on uh Forbes or different pieces that uh that can move uh the needle.
- 16:57SPEAKER_02: It's it's short-lived, so that's why you need consistent work and not just uh uh one-time play.
- 17:03SPEAKER_02: Uh, but there are all these different uh activities you can do to compound and over time uh improve how LMMs perceive you.
- 17:12SPEAKER_00: So you you can't just have agents go write a bunch of content and put it out there, right?
- 17:21SPEAKER_00: I mean that doesn't you can't game the system, or can you game the system?
- 17:24SPEAKER_00: Because search was gamed, right?
- 17:27SPEAKER_02: Uh so it's interesting because it's it's quite a similar phenomenon, like and and we're actually seeing this now uh uh in so up until like if you're looking this year, uh on just on ChatGPT, Reddit used to be a huge driver.
- 17:42SPEAKER_02: So citations of Reddit, ChatGPT would go see what users think, uh and and bring that and take that into calculation when it would recommend a brand.
- 17:52SPEAKER_02: And then the industry started gaming Reddit, putting a bunch of uh stuff there.
- 17:57SPEAKER_02: Uh so ChatGPT uh this year kind of did a few uh decreases, and in last August they they severely decreased uh the amount of citations.
- 18:06SPEAKER_02: So it's treating Reddit, but not as you know, a very important or prominent source, and we're seeing this in the amount of citations.
- 18:14SPEAKER_02: Uh so yeah, the system is being gained, but then the the you know the algorithms are are being changed so that it would be much more difficult and that you need to do the actual work of you know building the infrastructure, uh investing in uh in uh in persistent uh um communication, having a PR strategy, having social presence, being on LinkedIn, that kind of doing the the actual work rather than just as you mentioned, having agents.
- 18:42SPEAKER_02: Uh you can use AI, by the way, to help you with that, uh, but it needs to be from a more strategic uh standpoint, uh and and on a more consistent basis.
- 18:52SPEAKER_00: So so yeah, so what I'm getting is you you gotta have a strategy, you gotta be consistent, uh you gotta be in the channels, um, and you've got to sort of be monitoring how how how this evolves.
- 19:07SPEAKER_00: What do you what do you think is the important what where does um third party come into it?
- 19:14SPEAKER_00: Is it is is that becoming how are the because the always is obviously some somehow the LLM s was sort of thinking we can't we can't believe that we can't believe the bullshit that the company's saying we've got to have a we've got to have a balance, whatever check and balance, right?
- 19:30SPEAKER_00: So that's where they went to Reddit.
- 19:32SPEAKER_00: Um so I was thinking like when I was thinking about this myself, I was thinking of like these two dimensions.
- 19:45SPEAKER_00: One is sort of brand reputation.
- 19:47SPEAKER_00: Is it a good company?
- 19:49SPEAKER_00: Does it do good things?
- 19:50SPEAKER_00: Maybe even things that we sort of put to one side like sustainability actually become much more important than marketers today would have you believe.
- 20:01SPEAKER_00: Uh and then you have the SKU level, like so say take a company like Patagonia.
- 20:08SPEAKER_00: So Patagonia has great sustainability credentials, it has a lot of third-party endorsement, but then I want to buy a jacket.
- 20:15SPEAKER_00: So I've got two levels of the user journey.
- 20:17SPEAKER_00: I want a mighty looking at the reputation of Patagonia before I actually decide to buy a jacket, and then I go in and I'm looking for a SKU.
- 20:24SPEAKER_00: What I was wondering was there's actually sort of become does it does a does a whole new layer of um certification kind of emerge.
- 20:38SPEAKER_00: So people who go test jackets and people who is this kind of be could this be a new emerge or even old old institutions that used to do this, quality standards people, people that were trusted third parties, isn't it's now become a thing that you actually should be taking your products to these third parties?
- 20:58SPEAKER_00: And is that is that does that potentially have a competitive advantage?
- 21:04SPEAKER_02: Yeah.
- 21:05SPEAKER_02: Uh so I think something in my head.
- 21:07SPEAKER_00: I thought, you know, maybe this could be interesting.
- 21:08SPEAKER_02: No, that that that that's um that you you're thinking on it on the right direction.
- 21:13SPEAKER_02: I think what we're seeing is that with the decrease of uh of Reddit as such, there is an increase of these third-party commentatory.
- 21:23SPEAKER_02: Uh you know, every niche has the its kind of I don't know, gardener for you know specific uh software or cyber, the industry.
- 21:30SPEAKER_02: Uh so there are these websites which are more difficult, uh to you can't really game them, right?
- 21:36SPEAKER_02: They kind of give industry commentary.
- 21:38SPEAKER_02: Uh so those are are definitely ones which LMs are starting to go after.
- 21:42SPEAKER_02: And then also just review sites of uh websites.
- 21:45SPEAKER_02: So if you're a brand and uh on uh BBB or Trustpilot or even Google Maps, uh you have you know shitty reviews, then the LMs will see that and will say, okay, well, let's look at these reviews.
- 22:00SPEAKER_02: Okay, this is you know not a trustworthy brand.
- 22:02SPEAKER_02: Uh clients are not happy with its service.
- 22:05SPEAKER_02: Okay, then I'm I'm kind of discrediting uh that those person I'll I'll probably I'll probably not recommend it as often.
- 22:11SPEAKER_02: So I think it's kind of yes, investing in that uh uh third-party commentary.
- 22:18SPEAKER_02: Uh it can also be, by the way, like a PR strategy just to get your word out there.
- 22:23SPEAKER_02: Uh uh, but definitely invest in the in the places that the it matters for your industry, and also make sure that you the you the that you know you serve you serve your user well, the products are good, and that people hopefully speak about this uh online in a way which is uh uh good for your for your brand.
- 22:43SPEAKER_02: So I I think it's not anything new, I think, uh than previously, but uh, but I think the the in in the old age, I think, or in the search age, you know, enough links, enough kind of content you would get to to a top ranking on Google.
- 23:01SPEAKER_02: Uh and that would pretty much trump almost everything bad that you do.
- 23:05SPEAKER_02: And I think with BLMs, there's kind of less places to hide.
- 23:09SPEAKER_02: You really need to make sure that you have uh your uh 360 together.
- 23:15SPEAKER_02: Um it's not it doesn't mean that if you know you start at a lower basement uh baseline, you can't improve that, uh, but it takes a much broader and holistic and uh uh approach than I think previously.
- 23:27SPEAKER_00: Yeah, I was I was sort of thinking like you know, if you think about traditional marketing, a lot of it was about uh spin, you know, it was about creating a spin.
- 23:38SPEAKER_00: It was like okay, get this get this story out there that tells everyone that we're great and don't let don't let them look under the hood, you know?
- 23:49SPEAKER_00: And and so people could get away with a lot, you know, they could they could do a lot and no one no one questioned them because they didn't have the tools to question.
- 23:58SPEAKER_00: You know, I mean we you know we we went through the 1950s in America where doctors were telling people to smoke cigarettes.
- 24:07SPEAKER_00: Right?
- 24:07SPEAKER_00: I mean that's how bad it was.
- 24:10SPEAKER_00: And um so I I think it's I think this is a such an interesting dynamic for a business because business used to they used to be able to say how much do we have to pay to cover this up?
- 24:29SPEAKER_00: And now they might be in in a meeting and they have to say, actually, we can't cover it up.
- 24:36SPEAKER_00: We actually have to improve the product.
- 24:38SPEAKER_00: Because if we don't improve the product, we we will get found out.
- 24:42SPEAKER_00: And so I think that's fascinating.
- 24:44SPEAKER_00: That's a fascinating dynamic.
- 24:47SPEAKER_00: I think the temptation, I mean temptation is people will people will always businesses aren't perfect, they're kind of they're a mix of good and bad.
- 24:55SPEAKER_00: There are good bad things, you know.
- 24:57SPEAKER_00: Every business, even like even Patagonia, say you know, we're not perfect.
- 25:02SPEAKER_00: I mean that's why they're good, because they actually say they're not perfect, but every business has an issue somewhere that is discoverable and more discoverable than it's ever been.
- 25:14SPEAKER_00: So um yeah, I think that's uh I think that's an interesting question that you probably um face when you're dealing with your clients.
- 25:24SPEAKER_00: You know, it's like what what what what's what's like what's I mean advertising agencies are notoriously sort of naive in their um sort of assumptions that they can change the world through persuasion and that an ad campaign can make somebody feel like completely different about a brand, and you know, they sort of go in with that naivety, and but in your world, you're sort of telling your your diagnostic because you're sort of telling the client sort of what the world thinks truthfully, and I think that might be a quite a difficult thing to tell people.
- 26:03SPEAKER_02: Yeah, uh so we found that like most clients like if they have bad stuff under the carpet, they obviously know about it, but I agree that it's it's much more out there.
- 26:15SPEAKER_02: Uh and yeah, the the the the monitoring diagnostics definitely you you you can't hide.
- 26:23SPEAKER_02: Uh um so yeah, it does it does make you.
- 26:27SPEAKER_02: I mean, you can I think you can still get away with uh an okay product if you invest on uh but it it makes it so much more easier if you know you you actually give good service, if you actually give a good product, uh then it's more about okay, now how do I just get get people excited about it uh uh to share, and that's where you know influence for marketing suddenly can can influence that as well.
- 26:54SPEAKER_02: People post uh stuff on uh on on social and on YouTube.
- 26:58SPEAKER_02: Uh so I think that's uh uh it really kind of interconnects all the different pieces when you're thinking about the brand's presence, marketing.
- 27:08SPEAKER_02: Uh, I feel because uh VLMs consolidate everything into you know a single answer, uh, then then you need to make sure you you have all your ducks in a row.
- 27:22SPEAKER_00: What do you think about just coming back to sort of what we've been talking about, the the saying versus doing?
- 27:29SPEAKER_00: You know, you can you can you can say a lot of stuff, you can write a lot of articles, but is doing matter?
- 27:36SPEAKER_00: Does doing matter?
- 27:37SPEAKER_00: Does so when you're talking about a PR agency or something in your portfolio, you can have all the philosophical theories about what you want to do, and you can write those thought leadership pieces, but do you get extra credit when you prove that your thought leadership has been actioned?
- 27:55SPEAKER_00: So you have case studies with proof.
- 27:58SPEAKER_00: So I think there's these two different balances, you know, it's like one is having the thought leadership, but the other is I think you get scored higher, probably if you have consistent cases where you brought your theories into practice.
- 28:10SPEAKER_02: Yeah, and and that's where I think the the the the third-party review sites where you know customers review you and it it and and thinking about again big brands, and we're seeing that across different dimensions.
- 28:25SPEAKER_02: There are websites which rate agencies and and clients kind of can uh can rate them, and then there's there's stuff almost for every review site for every type of uh uh business.
- 28:39SPEAKER_02: Um that's where I think those come into play.
- 28:42SPEAKER_02: So um so LMs are definitely looking at that to understand what do consumers think about this product, about these brands, um, and and distill really the the type of stuff to that to your earlier point are not easily gained.
- 28:59SPEAKER_00: Yeah, but they have historically like Amazon has been famously his you know pointed out as like the bunch of paid reviews.
- 29:07SPEAKER_00: And but are they are they uh are you seeing the sophistication in the being able to in the LLMs being able to understand?
- 29:16SPEAKER_02: So they they yeah, so uh they work with what they have, but if we're thinking about the again the the the the business bureau BBB and trust pilots, just credible, there's credible reviews, then they know what it's like.
- 29:29SPEAKER_02: Yeah, exactly.
- 29:30SPEAKER_02: Um and and also how much scale.
- 29:32SPEAKER_02: So if you know you only have three good reviews, then maybe it wouldn't credit it.
- 29:37SPEAKER_02: But if there are thousands uh across again, even Google Maps or etc.
- 29:42SPEAKER_02: then that that that's more difficult.
- 29:44SPEAKER_02: Uh but yeah, I I think that they they they they tend to start looking at the the stuff um which requires so if we're talking about like in our industry there are uh places where the uh the website actually interviews clients and only then publishes their uh their uh review.
- 30:05SPEAKER_02: So when you look when you have any any LLM look across agencies and ask for agency recommendation, they increasingly look on those types of websites because they know that these are you know uh legit.
- 30:17SPEAKER_02: So I think they're they're kind of doing their uh due diligence in every niche in every market to understand, okay, what are the websites?
- 30:24SPEAKER_02: And you can see that by the way, it changes between LLMs.
- 30:27SPEAKER_02: So uh ChatGPT might go you know lean on on G2, whereas uh uh another uh Grok or uh uh or another LLM might pick its uh sources a bit different, and that's why it's important to have just a consistent strategy throughout and not just try and you know game Reddit, and now Reddit is not as important, so I'll move on to the next.
- 30:53SPEAKER_00: Yeah, no, no, that makes that makes sense.
- 30:56SPEAKER_00: Um and how like I remember talking to somebody pretty recently and they said these algorithms change with such rapid frequency that the you know you just you needed a it needs to be a dedicated effort to keep on top of it, right?
- 31:15SPEAKER_00: That uh it's like it these things are moving, they move and uh so so part of what you do is is offering the client the chance to basically track, you know, track their brand and um be able to see and respond.
- 31:34SPEAKER_00: So do you see the sort of becoming like a you talked earlier about agents and agents in I where I saw that in my mind was the way you're describing it, it was more like um reporting uh tool sets and you know doing a lot of the grunt work of bringing data together.
- 31:56SPEAKER_00: Um do you sort of see a can you envisage a moment where agents actually kind of optimize that based on so you take the reporting and it sort of becomes a strategic instruction and then the agents help um action those instructions?
- 32:16SPEAKER_00: Is that feasible or is that is that science fiction?
- 32:19SPEAKER_02: Yeah, no, I think I think that's that's completely feasible.
- 32:24SPEAKER_02: Um yeah, and and and I was thinking, and you can go to the other side where who would be the one actually you know searching uh and doing it would also be agents because the the consumer would say, hey, like find find me good set of running shoes, and would not need to go through the whole consumer journey.
- 32:42SPEAKER_02: So there's a whole uh machine to machine, machine to machine.
- 32:46SPEAKER_02: Yeah, exactly.
- 32:47SPEAKER_02: Of how do I make my website, my visibility like apt for uh uh for agents.
- 32:53SPEAKER_02: I think that's still you know, in even the AI search is new, but this is like really uh novel and new and still being um like the the all the different uh LLMs, Google came up with their web MCP, which is kind of uh a way for agents to interact with your website without actually needing to scroll through it.
- 33:14SPEAKER_02: So the the there is kind of the stack, the technological stack being written and built now to enable agents to actually take uh actions on on users' behalf.
- 33:25SPEAKER_02: Uh, but that's definitely where where things are heading.
- 33:29SPEAKER_00: So we're seeing like in B2B marketing, like at Salesforce, you probably got that happening right a lot.
- 33:35SPEAKER_00: You know, you've got people that got leads, they're sending content, the agents are identifying the leads, scoring the leads, sending the communication out.
- 33:43SPEAKER_00: That's kind of almost like an automated agent-based marketing system, right?
- 33:49SPEAKER_02: Yeah, I I those types of things are you know already existing uh today.
- 33:55SPEAKER_02: Uh I think most of them aren't as sophisticated uh in terms of you know actually taking on the uh the full correspondence or the full sales cycle.
- 34:06SPEAKER_02: Uh, but yeah, you can definitely have uh AI like do qualification, do nurturing, uh, get get the lead uh into into a point where it's ready to speak.
- 34:19SPEAKER_02: And there's I know of of companies that already have uh kind of an AI calling, a prospect, speaking through the the kind of just coordinating the demo.
- 34:31SPEAKER_02: The demo is run by a human, but but making sure to understand where you're from, what's kind of your where should I put you in terms of uh are you an enterprise, are you an SMB?
- 34:42SPEAKER_02: Which team should I yeah, exactly.
- 34:44SPEAKER_02: So that that's 100% agents are already doing that today in the more advanced companies.
- 34:50SPEAKER_00: Where where do you when you look out and you uh you said you have a uh corporate acquisition, someone who looks is looking for potential acquisitions?
- 34:59SPEAKER_00: What what do you see emerging as new specialties that you know are coming at resulting from this?
- 35:07SPEAKER_00: The sort of things that we yeah, the the thing what things we've described kind of like are fairly known, like right influencer marketing, PR.
- 35:16SPEAKER_00: But then do you think there's a new um a new world of new specialties that could be emerging?
- 35:24SPEAKER_02: Yeah, so I I think just the notion of uh we're we're and you talked about this a bit in the beginning, uh that there's such a gap, I think, between uh companies like us and just the marketing uh world, which I think by nature is very dynamic and and you need to kind of uh because all the the the uh companies around Google, Facebook are meta, are increasingly forcing uh the industry to to be more uh AI first.
- 35:55SPEAKER_02: Uh then you have we have all kinds of clients in you know more uh in other areas which haven't been yet as disrupted, and they barely do anything with AI, and that's uh uh I think an immense opportunity uh also to kind of help them with doing that AI transformation.
- 36:14SPEAKER_02: And so that that that's what kind of what we're excited about and looking at uh how can we help uh our clients building AI tools, AI automations, start to think about how to have um how to do AI in you know the finance in a regulated finance space, in a regulated health sector, uh uh which are usually the industries which are a bit more you know hesitant to to take on new technologies.
- 36:43SPEAKER_02: Um so that that that's the places we're um we're we're very excited about, and not I think just us, all the big uh uh Androtecov in AI are also doing stuff in that space because that's um what they need you know to continue to grow to get people to actually use uh AI and build up uh on AI, and and there's still a gap there with the the layman or the end, you know, uh in the the corporate uh uh the old corporate world until and and the the amount of use of AI that they're actually doing.
- 37:16SPEAKER_00: So so going back to yourselves for a second, um how much you know I wonder if this is instructive that you think about yourselves as a brand, right?
- 37:29SPEAKER_00: So, you know, you came, you were one thing, and now you're now you're so you've obviously probably you and your founding partners and everyone who's involved, your leadership have had to thought it think about what your brand is now, you know, what it was, but what it is now, and then what those principles are.
- 37:49SPEAKER_00: And I always think that this is a this isn't another emerging specialty, which is if machines are making decisions, uh on whose behalf are they making those decisions?
- 38:02SPEAKER_00: If that brand isn't encoded into something to do with the way that the AI thinks and behaves, then the AI is just going to optimize.
- 38:12SPEAKER_00: So if you are a luxury handbag maker and you're telling it to go procure uh leather, it may just decide to buy the cheapest leather and then the handbags you get made are terrible and they fall apart within three weeks and the brand gets brand gets destroyed.
- 38:27SPEAKER_00: That's my most extreme, extreme example.
- 38:30SPEAKER_00: But the idea is who's where is this who's doing this governance?
- 38:33SPEAKER_00: So even in your own company, when you're you're getting the agents to do the report, you want to make sure that your brand is part of that, and that's not like that's not just like a logo, it's like the way this thing is put together, the way it's written, the the perspective, the point of view.
- 38:50SPEAKER_00: Is that something you paid a lot of attention to or are thinking about or have thought about?
- 38:54SPEAKER_02: Yeah, 100%.
- 38:55SPEAKER_02: I think we um that that I I I love that uh leather example, but I think that that's where you need the the quality assurance and the the like evaluations, even if uh somewhere along the way you have another AI evaluating an AI uh A, we make sure that uh whatever goes out, like a client deliverable, always goes through uh QA of a human.
- 39:24SPEAKER_02: Um so that that's I think one level that we as an agency, it's obviously not uh uh doesn't apply to all industries.
- 39:33SPEAKER_02: Um but also having uh again, as I mentioned, evals for the different pieces, uh different uh uh uh places in the process uh to make sure that the you know quality is up to standard.
- 39:48SPEAKER_02: And and as you mentioned, uh you need to be much more explicit in the standards, in the rules, and what's acceptable and what what isn't acceptable.
- 39:59SPEAKER_02: Um and that's something that is is often something that you you build on.
- 40:04SPEAKER_02: Like we so thinking about kind of anthropic and clods, so there are certain you can write like rules that apply for the whole organization, what to write, what not to write, what to use, what not uh to use.
- 40:16SPEAKER_02: Um and what we are building internally is like uh so there's all these skills that people use to to do stuff uh on their behalf in automations.
- 40:26SPEAKER_02: Um we're building like uh uh uh guard rates, so to speak, so that everything would work in a specific environment, it would be monitored, it would be uh this is things that I think eventually all the LLMs would offer that.
- 40:43SPEAKER_02: It's just still being built.
- 40:45SPEAKER_02: Uh, but 100% it's not like the the you know, it's like I think Sam Altman gave the example that he wouldn't tell uh an LLM, hey, like if you could just go over my emails and respond and like be done with it, uh then obviously there there's big risk there, assuming you know you're you're you're a person that matters and your emails matter.
- 41:06SPEAKER_02: Uh similarly, like just going and buying.
- 41:08SPEAKER_02: But if the context is okay, you're an agent that does that is in charge of optimizing this campaign, uh, and you know, tell them and and then give me daily report on what you did, what not you did, and set certain guardrails on what you can do or when you need to ask me, that's and and this is all this agent can do in terms of the amount of uh permissions you give them, etc.
- 41:32SPEAKER_02: That that's I think the way to to think about it, to have the these uh uh very kind of uh uh regulated, governed uh places where the the actions are limited, you're aware of kind of the the risk, um and and obviously have you know monitors in place to make sure things don't go out of whack.
- 41:55SPEAKER_02: But uh going A to Z in terms of automating your whole life process business is um is is is something that you need to take.
- 42:06SPEAKER_02: You need to have many stop signs uh on the way because that these are you know essentially uh things that are make a lot of these tiny decisions of logic that we have uh in our company in as humans um and are are implicit, need to be much more explicit when dealing in the AI world.
- 42:29SPEAKER_00: Yeah, which which is which is really fascinating because you when you are talking about like um taking kind of legacy or old companies who have AI who are kind of at the early stage, you look at something like Palantir and like the FD the forward to play engineer, um, and there's that ontology, that that whole thing is they're kind of building a schema of like how the company makes decisions and they're kind of setting a ceiling and where you know where the where the automation stops and where a human takes over.
- 43:04SPEAKER_00: It seems and and so much to me of that is like it, you know, you could be working with a bank and you could be looking at their brand guidelines, and they could say, for 25 years, this bank has been dedicated to improving the lives of young people and helping people get their first homes.
- 43:21SPEAKER_00: You can you can say that in very woolly marketing language, and no one's gonna ask you for specifics.
- 43:28SPEAKER_00: But when you come into this world, they're saying, what do you mean by a young family?
- 43:33SPEAKER_00: How how so the specific people are gonna I what I'm saying is there's so much more rigor as to your belief system and how that belief system drives decisions that is sort of potentially transformative, right?
- 43:53SPEAKER_00: You know, because um people have never been asked these questions before.
- 44:00SPEAKER_00: They've never had to.
- 44:02SPEAKER_00: And so I think that's a that's radical.
- 44:05SPEAKER_00: It's it's radical and it's extremely complicated.
- 44:10SPEAKER_00: It's not simple to do.
- 44:12SPEAKER_00: Um so I don't know, I don't know if if if you I I don't know how you think about that.
- 44:20SPEAKER_00: It's is it it seems to me that you've got to start very small.
- 44:27SPEAKER_00: That that that you have to especially with these companies, that you have to take a little piece, you have to take something that you can experiment with, prove to them that it works before you move on, because it's just too complicated.
- 44:44SPEAKER_02: Yeah, I think it's any kind of process.
- 44:48SPEAKER_02: Uh yeah, that's where the what I mentioned earlier on on breaking uh a process into kind of agents and sub-agents or even subautomations where you have tight control evaluations.
- 45:02SPEAKER_02: Um yeah, you you you have to.
- 45:05SPEAKER_02: I mean, I wouldn't I would also be careful of starting too small, uh, because what we found when we started kind of our our uh AI journey is that we we we had we started by mapping different processes, etc.
- 45:19SPEAKER_02: And we found that people are you know automating uh bits and pieces of the grunt work and taking uh data from this sheet to that sheet rather than looking you know on the whole parse process at large.
- 45:33SPEAKER_02: Um so I I think there is room to kind of rethink.
- 45:38SPEAKER_02: Do we even need to do that part?
- 45:41SPEAKER_02: Given you know we're we're rethinking everything and what and and that you do have uh you are able to add reasoning.
- 45:49SPEAKER_02: Uh so I wouldn't go like too small because you need to rethink uh how how you're how you're doing stuff, but definitely not try and uh uh and do everything at once um because the risk there is is too high.
- 46:04SPEAKER_02: So I agree with you that you need to chunk it into different processes and and uh and try to think about this like what's the output of the first process, what's uh uh evaluation, what's the QA, and what's acceptable to move on to the next.
- 46:20SPEAKER_00: So a big part of this has gotta be like there's gotta be a huge training element here.
- 46:26SPEAKER_00: It has to be, right?
- 46:28SPEAKER_00: That has to be a huge part of the business because um you're taking someone who has a perspective of a function, and then you're taking workflows and you're automating those workflows.
- 46:42SPEAKER_00: So that's like okay, there's a lot of going on there, right?
- 46:46SPEAKER_00: You're taking you're looking at someone's job and you're telling them that the job that they used to do is becoming quite different.
- 46:55SPEAKER_02: Yeah, I I I heard uh um another podcast, and they were speaking about the fact that how they're trying to hire more people with uh uh that are T-shaped in terms of their background.
- 47:08SPEAKER_02: They have a specialty, but they also have a wide uh background.
- 47:12SPEAKER_02: So they can go deep on one niche, but but they there it's not like that that's the only world that they live in.
- 47:17SPEAKER_02: And I think this is what the the the this new era calls for.
- 47:21SPEAKER_02: You can't be just you know this tiny uh uh specific thing, because the the the the positions change, the roles change, and the the you have to have uh broader context on you know your part in the broader uh company in the broader kind of motion.
- 47:42SPEAKER_02: Uh uh because exactly as you mentioned, the the the AI by nature would would shift the processes and you'd need to suddenly uh be able to QA or or or give it instructions for stuff that maybe wasn't under uh your previous role.
- 48:00SPEAKER_02: Um so yeah, part of it is is is kind of doing the upscaling of the team, and we've built internally an AI Academy towards that.
- 48:09SPEAKER_02: We're investing a lot.
- 48:10SPEAKER_02: Um, but I think it's also people that have a high sense of agency uh and again have that um T-shaped uh background, and and and it's a my I think it's a mindset uh shift, and uh those that go through it and kind of uh do take the the the the pillow so to speak, are we see that employees kind of that uh uh from where they when we just when we just started off to AI to today, and the amount of output, the productivity, the amount of impact they have is you know uh uh 10x.
- 48:46SPEAKER_02: Uh so I think that this is what this era calls.
- 48:49SPEAKER_02: These were the ones which will uh uh you know be the the employees that companies will go after.
- 48:57SPEAKER_00: And that's really really interesting.
- 48:58SPEAKER_00: Because I think another like another model I had in my head was uh you know, take a take a very good financial analyst working in an investment bank.
- 49:10SPEAKER_00: What I was kind of thinking is if so much of their work the anal and the financial analysis can be done for them, what does it mean they now do?
- 49:24SPEAKER_00: It's almost like because there is no like there is no sort of uh it's a process.
- 49:32SPEAKER_00: You know, every maybe um maybe there's a slightly slight difference in perspective that uh a JP Morgan report has versus a Goldman Sachs, but ostensibly they're pretty much the same kind of format, they can have the same things.
- 49:49SPEAKER_00: So if the analyst is now freed up and has more time, what do they do with that time?
- 49:55SPEAKER_00: And we don't really know that, right?
- 49:58SPEAKER_00: Because we've never really done that before.
- 50:00SPEAKER_00: So it goes back to who's in that role and who has the aptitude to basically maximize their value add given that they've got the time, right?
- 50:13SPEAKER_00: So it's it's sort of like a it's I I was thinking maybe it's like what's gonna happen is somebody, these people will go to their bosses and say, I've got this gr I now have this free time.
- 50:30SPEAKER_00: I've been really thinking, I've done my reports analysis.
- 50:34SPEAKER_00: I'm I'm an expert in legislation.
- 50:36SPEAKER_00: I know way more about legislation of like energy.
- 50:40SPEAKER_00: I want to set up this unit to generate more money for the bank because I have this expertise, and the bank will either say, yes, you can do it, or they'll go and leave and set up their own thing.
- 50:54SPEAKER_00: So I think this is kind of really interesting dynamic of like back to those T-shape, back to those sort of almost I think entrepreneurial, entrepreneurial individuals who kind of see that the technology is an assistant that can help them redefine what they do.
- 51:16SPEAKER_00: And yeah, you're right.
- 51:17SPEAKER_00: I think it's so interesting that the best use cases are the people who are using AI as a superpower and don't see it as a threat.
- 51:30SPEAKER_02: Yeah, uh, and it's it's a mind shift.
- 51:33SPEAKER_02: I I heard that uh the I read the a piece by Anthropic and they talked about how you know the bottleneck used to be about uh like building the actual uh products or or code.
- 51:47SPEAKER_02: So now AI does that, so the bottleneck moved to reviewing the code, but now AI does that.
- 51:53SPEAKER_02: So now that now they're kind of a stage of what to build.
- 51:57SPEAKER_02: So that's where the kind of the human element is uh to choose just what to build, and then the AI builds it, it reviews it, it uh uh but what happens when the the AI will also take over that part.
- 52:09SPEAKER_02: It's it's very interesting as a thought process to think kind of where where the where the next bottleneck of you know uh productivity is.
- 52:17SPEAKER_02: And I think your your example was spot on because I was thinking of how does this apply uh outside of the realm of uh of uh you know coding and building products and uh so other verticals, other business verticals, other um uh initiatives.
- 52:34SPEAKER_02: Um yeah, very very interesting times we'll be at.
- 52:39SPEAKER_00: Cool.
- 52:39SPEAKER_00: Any um thoughts, advice for people who are concerned about this or feel that they don't know enough, or what what what what should they be doing?
- 52:55SPEAKER_02: Um so first I I I think just just be doing.
- 53:00SPEAKER_02: I think that that's just take the first step.
- 53:02SPEAKER_02: Uh like try it out.
- 53:04SPEAKER_02: And if you are not paying for uh the uh um uh the the premium seat on any of the LMs, then spend those$20 so you can really experience kind of the frontier uh models.
- 53:16SPEAKER_02: I think those are becoming table stakes today.
- 53:19SPEAKER_02: Uh and it's it's I think once you you you're in, you kind of especially since it's so easy because you can ask anything and suddenly you have some thing on the other side that can really take you step by step no matter where you uh uh where you're starting.
- 53:34SPEAKER_02: Uh and I think on the on the company level, as we did as we discussed, just start by knowing where where you are.
- 53:41SPEAKER_02: Uh go to uh NAA report, you can check ours out, uh, and and at least know where you're at and where your gaps are so you can start um actioning in.
- 53:52SPEAKER_00: Brilliant.
- 53:53SPEAKER_00: Thank you so much for your time.
- 53:54SPEAKER_00: Great conversation.
- 53:56SPEAKER_00: Have a good rest of your day.
- 53:57SPEAKER_00: Okay, man.
- 53:58SPEAKER_00: Bye.
- 53:59SPEAKER_02: Bye.