How to actually get your brand into AI search results | Guillaume Ang @Psyke

Speaker A: Foreign. Hey there. Welcome to another episode of Sauce and Bound. And today with me is Guillaume. Welcome to the show.

Speaker B: Hi Anna. Great to be on the show today.

Speaker A: Awesome. Well, it's great to see you. I think this is going to be a super interesting conversation because you guys are doing something that a lot of people want to do or are trying to do. And as a marketer, I am a little bit afraid to do, I'll be honest. But maybe today is the day when you convince me that it's all good. Before we jump into this, maybe just a bit of an intro who you are and what is Psych.

Speaker B: So I'm Guillaume, very obviously French, been in Australia for the past 13 years though and doing startup for the past, I think 17 years. My journey has been from an engineering background, turned into a business person and I've always navigated at the edge of the two worlds. I did, I think my first exit at the age of 27 in the space of B2B SaaS for chemical industries regulation. Right. And that was in Europe. Left for Australia, wanted to experiment in an uh, English speaking world. Tried to improve my English accents. Didn't succeed at that. I ended up having a blast, kind of launching one of the first groceries delivery startup in Australia. Two hours groceries delivery on the Uber model. We ended up exiting to a local player which was a very exciting five years journey. And then I started to do investment myself, help other startups, founders. I actually ended up creating a consultancy in the space of using low code and no code tools for growth and help local players grow that way for a few years. And on um, the back of that I felt ready to go back in the startup world. And there was this growing realization that came to me as an advisor and investor that all the founders and managers they were working with had this same kind of recurring pain about marketing that came with so many different channels. Each channel's being very technical and I was seeing a lot of overwhelmed people and I was seeing a lot of marketing managers not being able to get their goals down just because of the sheer complexity of the space. And look like this hasn't improved. Like marketing is harder than ever. But what ended up happening is that with my mentor at the time, Matt Brown, today a co founder advisor of the business, we decided to solve that for good. And that was six years ago. And we launched what was called then Upfluid and then as we named Psych last year, with a clear mission to help innovation spread faster, to help innovators focus on what they were great at and really shorten the feedback loop between kind of having an insight to getting traction and getting learnings from a marketing perspective. And we launched the first product, then AI came around, we raised a fair bit of money to accomplish this vision. And we've been hard at work for the past two years, really helping marketers solve one key thing first, which was their SEO and GEO presence, which is the kind of the AI search visibility. I think we're pioneering a lot of technology to create more pages in a safe way to educate LLM to how to understand your business. And we ended up creating this growth engine that's able to train the machine out there and the LLM and Google to how they should understand your business and who they should talk about your business about. And now we're expanding wider horizons as we realize that we can help with more gross organic engines. And yes, it's been a very exciting journey and, um, looking forward to answering any question you might have about it.

Speaker A: Awesome. Um, yeah, it seems like you've done it all. And to be honest, from all the recent conversations with founders, everyone says people who have both marketing and engineering in their wheelhouse, they are the one who are winning. Because as a marketer, honestly, I've been pushed quite hard into learning how to orchestrate marketing and how to use tools like cloud code nhten, blah, blah, blah. Like there are so many of them now. And at the same time, tech founders have really been unleashed, right? Like people who have never wrote a marketing copy now think, okay, we can just program that machine, they can give this to us. And I guess that's exactly what you're doing. But I can't help but wonder, as always, when someone promises, hey, we're going to create a thousand pages about your product. It was five years ago the concern that Google is going to penalize you for this. It still is a growing concern if Cloud ChatGPT or whatnot are going to penalize you in the future. So I'm sure this is one of the things that customers ask you. So how do you deal with that?

Speaker B: 100%. And I think it's a great start to explaining and debanking some methodology around the space. The first thing I'm going to start by saying that we're launching thousands of pages every day live now, right? And so I think we are at a vantage point to look at empirically whether, like, what we've done is detrimental. And the overwhelming answer is no. Right. And what we see in the opposite side of that, we see returns in terms of lead and contract Signed versus the money we have spent with us that are in the 6-10x range after a number of months. Which proves to me that kind of whatever we're doing is highly scalable. If there was any takeaway from anyone listening to us today that whether you use someone like us to do it or whether you try to do it yourself, like it's really worth investments. And I think looking back 10, 20 years in the past, like the largest companies have always had not only hundreds of pages, not only thousands of pages, have had 10 of thousands of pages or hundreds of thousands of pages live at any one time. And so what we used to call programmatic SEO or long tail SEO has always worked, right? It has always worked. It has always been very expensive and used a lot of brain work to make it work well. And there uh, are very strict conditions for it to perform and not be penalized. It's always been a winning recipe for the larger business. And so there is actually like the one reason why small businesses have been afraid of getting to it is more from a mythology perspective. And sometimes I will say agencies not being able to deliver it properly and therefore down playing it in a way that was like it's not for you, but it's for the bigger guys that are making ton of money out of it. So I really want to state the truth that this has been in the space and if you actually look at the Myers, the Laurel, all the big side of it, like you're going to find hundreds of thousands of pages finder, right? So now what does it take to make it work? Well and now I want to take a step back and look at how LLM works to really I think offer a different angle on what will make it valuable and non dangerous. So an LLM ingests a lot of information from everywhere. It gets trained on this information. The training gets updated every week, every month, every quarter and the model evolves to adapt to the reality. A very interesting fact is that 70% of the content that is talked about by LLM of cited is less than 10 months old. So there is a constant training happening. Now if your website has uh, a very limited range of content, meaning like it's only talking about itself and some related case studies, right? Then you delegate to the LLM and to Google the load to abstract your content to specific use cases from the users. Now we need to understand that each user comes with a very heavy context. Now when you use ChatGPT as hundreds if not thousands of conversations as a story they can use to understand New specific contexts. So when the LLM m tried to match a specific request or query to a specific company, they need to abstract that. Now the right way of doing it right is by helping LLM have an easier time, spend less energy and compute power into matching the right query and the right context to your company, right? And for that it's completely legit and working very well, very efficiently to understand all the multiple contexts your customers might come from with an intent to purchase from you and create the right content for them that connects to the company. And in doing so, what you need to do for the alms is make a very easy job to package an answer that makes sense to them and help them deliver legit answers to their customer query instead of extrapolating content they don't have. And so the role of creating these thousands of pages is actually to ease the LLM into understanding how your brand fits specific queries. And a brand that does that versus a brand that doesn't is always going to win. And so if you're not doing it because you're afraid of too much content online, then you basically at fault and you're losing against your competitors that actually embrac the way LLMs and Google has been scraping the web. Now one last thing, it doesn't mean that you should create thousands of pages at scale that is as long because if you don't add any value to the LLM and if you regurgitate what the LLM already knows, then you basically creating something that's nice, that's time consuming for the LLM to treat without adding value to the training. And so the one way of doing it right is at the crossroads of three things. Your brand, what your brand knows about customers and what it stands for. Unique content that you will have. And this is a key ingredient if you have metadata m about your customers, about your industry, about your productivity gain, about anything that you know that specific methodology processes that is not exactly there in the space to train specific for specific context, then use that. And the third uh, thing is what the LLM m already knows. So the LLM only plays a minor role in creating the right content for at scale pages. And this is a unique recipe, right? And once you create unique value for that, then LLM are going to have so much better time and less compute to match your company to the right customers. And this approach of having unique value has been true for 20 years. And whatever the Google algorithm evolution, it's always been about that and it's always what they've said they will Prioritize. And I think it's more true than ever.

Speaker A: Right, okay, so I want to ask you exactly about that, right? Because there are millions of pages of AI slop right now, right? People just say, create this and that for me with no guardrails, with no opinion, nothing behind it, and absolutely no value. Like you said, it's just working and reworking the same information that's already there. Maybe a hack for a very short time, but doesn't add any value. And as a marketer, again, I'm thinking, where do we stand? How do we create content that is both very easily digestible by AI? Because we all think about AI visibility now, right? We want our content to show up and our brands to show up in AI search engines. But also, and some people may argue here that, oh, nobody reads long content anymore. But on the other hand, I still believe people want to have your tone of voice, your brand's qualities and values firmly, very opinionated in the piece that they're reading. So do you have any, like, practical tips? Okay. If you're starting, especially if you're, let's say, not a senior content person, right, who maybe still learning of what taste looks like or what good content that really ranks looks like, how do you prompt LLM to create something like that?

Speaker B: I would say there are a few phases. If you're. If the Persona you describe is someone that is unexperienced and as clothes, I think there is an easy from where you are to 70% of the way that's doable in two hours. And this is just like basic health tech that now every LLM knows how to do. Well, basically for our sake of what a good website looks like. Help me upgrade my website and I think most people will get there. The next layer is where it becomes, I think, way harder and where it doesn't work well if you're doing that on your own without the right preparation work. Right. So you can kind of throw things at the wall and hope that sticks. Right. And a number of founders I've met have actually got from M 70% to 80% with that. And that means, you know, like creating AI log and extending the surface area. All that they've been doing, actually doing that is, yeah, extending the visibility. It doesn't make them legit, it doesn't make them accurate. Now, the last 20% is kind of the process I will explain to you, which is, uh, imagine a Venn diagram where one circle is where you ingest your brand. And as you say, unique tone of voice, unique text on topics unique beliefs that I will incentivize every person going to that spend a good hour every week identify in your company what do you know that is unique and that can come in so much different ways. I'm um, going to give you three examples for the first thing like look at your customer tickets if you have a lot of customer or QC support tickets. Take this database, look at what people ask in terms of questions, look at what a successful answer looks like. So that's unique data because it's an interaction where you actually solve someone problem and it's an amazing source of that. The next thing is Salesforce recording. Give your sales calls recording to your AI, create this unique database of what's unique about your company. Another example is if you have a lot of customers on a product LED products then you probably have amazing metadata. Uh, you can abstract per industry, per location, per use case per case today and create unique data layers that basically you can give to the AI cefi. This is your unique data bucket. Now you're going to ask and the prompting is way more complicated than like in a sense you're going to ask the LLM to cross what they know about the queries, which is something you need to get from AHREFS or semrush. Just having a level data uh, with what you think your customer looking for and to cross this with the unique data that you have and find the opportunities to talk about specific things. So that's creating kind of your map of unique knowledge that you own that is going to be valuable to one person in the market. And so that says the preparation work. Now what you want to do with that is that you want to prompt the LLM to create pages that are performing well. And in that there is a number of techniques that ah, are not necessarily mastered by the existing skills in the market and you might miss them out. But like there are things like obviously having the markup done for your page, making sure your page is very intelligible for an AI, but also what we call chunking. And chunking is a technique where you prompt the LLM to get your content, the one you determine was very relevant to a specific context, a specific person with a unique data point. And you bundle that into about 400 words that talk about why, like what is the issue you're trying to answer, what's the context, what's the unique angle that you're bringing into it and what's your conclusion about it. And what we find is that when you've done this whole process well the chunks that you've created in your page and you can have multiple chunks per page are, ah, almost going to be printed as they are in your page when they get cited. Because you were able to influence that. Well, the training of the LLM. And so I think for me this is not gold standard, but it's like silver standard into what can be done. And I think if you follow the process well, you can achieve amazing results with it.

Speaker A: Awesome. Um, all right, that's very practical. That's what I'm going to play with after this episode. But also, obviously we've all been there, right? We've all prompted whatever LLM we're using. And then it returned with something and we were like, wait, no, that's not true. And it says, oh, yeah, you're right, not true. So hallucinations are happening. And it's not like, I think I told you about this very simple example because I'm going through a course where it is about the muscle structure. There should be no variations, right? It's the muscles, we all know what they are. And I was running the test that I did previously through LLM just to double check, to make sure I feel comfortable with it. And it made several mistakes on seemingly very straightforward information that is just out there. And you know, it's not a growth hack that everyone's got a different one. And of course it will offer you different answers and will hallucinate maybe sometimes. This was something very super straightforward. And I thought, wait, so, okay, so if I'm getting these very inaccurate answers for something that simple, how can I trust LLM to create a thousand pages? Like, how am I going to make sure I can then go and review it? And that brings me back to something I discussed with the, uh, developer team very recently, where they said LLMs create such huge chunks of code these days that you're basically just a reviewer and you have to go back and have this huge cognitive load of reviewing the work. So eventually, as a marketer who created a thousand pages, how relevant is it to think that, okay, maybe 30% of it will be hallucinations and uh, I will have to create guardrails or more agents on top of that will be checking for that.

Speaker B: It's a very fair question and something we've had to work around and work with. And so I'm going to share a few learnings. That is not the complete view of the world. It's what worked for us. And I think we're pretty proud of where we've landed with that. And I Think everyone needs to experiment with their own systems. But overall, if you think about agentic work, this issue is meant to arise in the same way that you're going to have employees. If you think about your workers like your AI workers as employees, you do have the same feedback loop. And if I look at what it looks like to train a great employee to do specific things, it's always going through the same thing which is like this is your frame of reference, this is what you need to learn. I now go through it, make mistakes, come back to me, let me in control so that I can fix mistakes. And so that's kind of delegation system that I think everyone going through training of agents, like training employees have to understand and experience. I recently published an article on my LinkedIn about the frustration of training AI employees slash agents and it's one of the best posted. Like I think the number of pressure went through the roof just because people were empathizing with something that not a lot of people say, which is that training agent is painful. But also the fact that nobody speaks about the frustration creates this expectation that training agents should be seamless and that you could get all your lead pipeline generated from it overnight. It's not true. Like training agent is painful. Like it is training great employees. I think that I will start there, don't expect it to work overnight. Now there is a pattern that I think has been talked a lot about and if you spend a fair amount of time doing agent development that you're going to bump into like very fast. It's like the model of creator verifier, right? And so this model means a very important thing which is that you actually instead of having one person in charge of delivering and you are the reviewer, you add filters and layers of that where an entity that is absolutely unrelated it to the first one is in charge of assessing the content. And I think you enter this pattern of like for me, like brain dissociation where you can have the same agent, but if one of them is incentivized to please you by creating great content, but the other one is incentivized to please you by figuring out that this content is bullshit, then they will have an interaction that will be very interesting and it's not a uh, one off. Like this needs to be a loop. And so this is where the use case for having like correlated agents checking each other work that iterate on each other to a point of stability is what you want. And if I was to go a bit kind of philosophical about it, I think this is One of the key things that agents have introduced versus skills, that you end up having agents with different personalities. And I think a point of equilibrium happens when people with different personalities tend to agree. And so for that you need iteration and you need people that have very different incentives in the same room. The same way that you reach better decision when your full exec team is around you and you have the product person and the salesperson and the finance person agreeing, you need to see it that way. And I think building that inside cloud is not easy. And so that's why I think there's still an edge for businesses to work with expertise. I can only hope that these kind of processes to help marketers get to the right equilibrium points are going to be democratized. And that's what we're working towards as well.

Speaker A: Yeah, it's funny you started talking about different personalities of agents, and I actually have a very funny story about that because I'm that person that always says, can you please give me this? Thank you. So I'm very polite with my LLMs, but my partner is always very down to earth, like, okay, give me this, give me that. And so eventually his agent is talking to him the same way. Like, and when I tried it and it was like, okay, yes, all right, let's get to the bottom of this. Let's please figure out this and that. Thank you. And the agent was kind of like, what's wrong with you? What happened here? Is it you? Are you okay? How are you feeling? I was like, man, okay. It brings me back to this article. For the love of Go. I can't remember who it was, but he was talking about the fact that LLMs might be conscious. Right. And we have to talk to them as if they're conscious. And I don't know, I'm a little biased, but, uh, where do you stand on this? This episode is sponsored by Reworkful. Looking for new ways to find customers for your SaaS business? Consider building an affiliate program. Rewardful is the easiest affiliate tracking platform to set up, manage, and scale. For SaaS companies, building a successful affiliate program can be a little bit intimidating at first. And that's why Rewardful has taken what they've observed from their most successful customers, affiliate programs, and distilled that into an exclusive online course. The exciting part, their affiliate marketing course is absolutely free. Start the course at academy.com rewardful.com academy.rewardful.com and turn your biggest fans into your best marketers.

Speaker B: Unfortunately. Unfortunately. Right. I think for me, m. As long as you can Turn off something. The level of control you have on it is absolute and therefore it's denial of freedom. So as long as we there you have a big influence on it. As you say, like the way you train them has an impact on the way they communicate. We have this example of an agent that was an orchestrator managing a new agent that was doing something. The new uh, agent had very specific goals, right. And the orchestrator was pushing on it very hard. Did you do that? And one of the goals the new agent had as a new agent was like you shouldn't push anything without supervision because we want you to gain trust from us. Right. And the other side, like the orchestra, like oh, you have goals you need to achieve. Right. The new agent became very shy and became very timid or super careful, almost a bit traumatized. Like you could read into that some level of kind of personality and human like behavior. The truth is we provide did a tool to contract story instructions and we had a tool in the middle and when trying to resolve which to give more weight it was getting stuck. And then I think the LLM was great at ah, expressing this contradiction through emotion. I'm still able to see the cogs for now. Now say there is something that was unplegable. Now that's a different question because the fact that I don't have control on this life and goals and mission, we change things drastically.

Speaker A: Yeah, absolutely, yeah. I mean this is more of a philosophical kind of part of this conversation I guess. But I'm just curious, like where do people stand because it's such a new thing and um, such a disturbance to many, many processes that we have that

Speaker B: I think there is maybe something I haven't heard said before, but I think for me is really true is that while AI itself is not conscious, the entity human plus AI is very much something new. And now you have like, like I wouldn't say my kids because they're six years old, right? But like if I take a 15 or 18 years old that has access to AI, they're going to think in a very different manner than we used to think when we were 18 years old. And so this entity as human plus AI is where you definitely an influence, right. And this is where if entropy or uh, ChatGPT releases something that becomes very fundamental to how people think every day, that's going to have a fundamental influence of the way intelligence is being used and manifest in the real world. So I think this is absolutely already live the pairing of human and AI. AI is something that has been impacted Already. And that has, in that way, consciousness.

Speaker A: Okay, that makes great sense.

Speaker B: Okay.

Speaker A: I want to go back to the company, right? Because you guys, ever since I started with SaaS, pivot has always been this feared thing, right? If you pivot, something went wrong or how do you survive and how do you make sure your customers survive with you? So you went through two, and I wonder how you do that. Plus the rebrand. So how to survive rebrand plus pivot in today's age. And how to communicate it not only with your audience, like I said, with your customers, but in your case, also with your investors.

Speaker B: Yeah, great question. Just to give more depth to what was a pivot? And there was mostly one. And then the rebrand came on the back of that. When in 2020, we started to work on the space of marketing software as a service, especially low code automation was an amazing impact on the industry. And so we were looking at places in marketing where no code could help, and we found one. And over a matter of months, we got 25,000 users, talks to the reception of this type of tools in the public. And so there was a path where we could have tripled down on this path. We had our database of users, we knew exactly what they wanted. And then in the same time, you know, the team started to play with AI and we had this intuition that within years, everything we built as a platform will be achieved in a matter of hours. And I'm sorry to say, today it's true. Like the tool we had in the past, we could actually create it in a matter of, of days today, which doesn't prevent it from being valuable. But in our spirit of really trying to deliver outcome to people, we felt that we were never going to land in a place where we were delivering real strong outcome on a long period of time. And so to align ourselves with that was for more of a founder choice. I think both the founders were exited founders, and we didn't want to build something that was short term, um, or not delivering expectation, or was betting on people not being like, technologically advanced enough to benefit from it. And I know a number of founders that we say, no, no, no, if that can make money, you should have done it. And it's a good point. We are very proud and happy. That came, became a realization. We had this very hard conversation for a few months to evaluate what it should be then, and that we ended up having the support for me invested. And I'm not saying that was an easy process. I think there was a big decision to reinvest towards Something new. But I think we were supported by one the deep knowledge we had of the space. And we had like hundreds of conversations supporting our assumption, but also a very strong sense that the industry was moving forward in a completely different place. And if I take only one key outcome is that we started then working on programmatic SEO at scale because we knew that was something that was unsolved, that AI could help us solve. And within months we realized that was also key for geo and AI search. We've actually achieved more revenue in one year than we did in three. I think it paid off. And you know, our investors called us like the dark horse of the the portfolio. And you look back and it's like you're very happy about that. But yeah, I think it was thanks to a very solid founding team plus investors community that were absolutely supportive. We're uh, thankful for that. I didn't like the hard conversation, but I don't regret them. Now to the pivot, like to the name itself, I think it was a different and I think changing names is actually a good thing because once you get enough feedback on the name and what people associate to it, and I think the previous name wasn't at all what we wanted to be associated with. And so that provided an opportunity for on something new. And also I think part of the new name psych is obviously something very linked to the human psyche. And that was something that is actually very dear to us that we didn't get to add to the name in the first iteration. And yeah, uh, we really believe that marketing is artist work of creating narratives that resonate for humans. And putting the human at the center of our work has been something very inspiring and something we will keep trending forwards as it gets m more place into our business.

Speaker A: All Ah, right. What I'm going to say, I mean in a bad way or kind of of downgrading what you guys did. But I think it's also in your particular example, I mean we say, you know, there is no recipe for success. Right. There is no real playbook. It's a lot of grind, a lot of hard work, but it's also being able to see the trend and to flow with it. And in your case, I think that's also something that they really helped. And a lot of founders struggle with that kind of natural or market uplift. You are still going into the unknown of how the audience is going to take what's happening. So. Super interesting. I wanted to ask you about rebrand, how it affected you in your visibility online and how you are dealing with that because I've actually had one team here on the podcast this year that went through a rebrand and they said it tanked like overnight on Google and since then they've been trying to climb out of this. What has happened to you and how are you tackling this? This?

Speaker B: Very good point. So the first thing is that I think the more search visibility goes, the more like the notion of entity is important. Some marketers would argue, uh, that entity SEO has been around forever. By that I mean that you cannot hope to sell different things and keep reputation for it. Right. In some ways I think it's important that people understand that. So if you're launching something that's quite different, you should expect to have a ramp up period for us. I think the thing that will impact from an SEO perspective I, uh, reprint the most is the domain name and the fact that now you need to reissue it. So you're going to keep a certain level of authority, but probably because the old links were kind of unrelated to what you're selling now, especially if you're selling something new, then you're facing the risk of kind of these links not being very valuable to your domain authority and so you need to take the hit. Like a funny story for us is that today our domain name is psy co P S Y K E dot co and because that's what we could buy all the time, it turns out that today we have the.com, we have sai.com, which will obviously look better. One reason that we're not activating it is 100% such visibility, this kind of side consequences and then we're going to be very happy with site.com for the time being. I think that's where most of the hits happen now. It's also a great opportunity knowing that LLM M, uh, have such a lesser time to refresh than Google that actually building a strong brand now takes less time as long as you're being very intentional, namely you can get citations faster than you can can get seen on Google search, especially if you start from like deep query perspective. So instead of trying to be against the main guy in the industry, start long tail, start small, start creating a niche with unique data, uh, for specific queries and then bring up your brand from that. And obviously do the digital pr, do the social media, uh, do the Reddit, do the G2, do everything that I think it takes for LLM to understand where you stand for and you also have the opportunity to articulate your change somewhere in your website site to LLM M to understand where you are now, uh, and why kind of part of your legitimacy is transferred to you. So it's a more natural language approach to this transformation rather than your data driven SEO that the thing was before.

Speaker A: Yeah. All right. On this podcast we almost exclusively talk about how AI is making SaaS better or us better or us more capable or whatever. And I want to ask you what's your take on what AI is still not good at and what you wouldn't give to AI just yet in your company or in your life?

Speaker B: I think the first answer for me is almost obvious, but I'm not sure it's a common answer, which is that, uh, the one thing that I cannot delegate to AI today is accountability. The first line in the center that I draw in my company is who is accountable. And if the answer is I've run an agent and is supposed to do that, then I know that accountability doesn't exist. And if it matters, I should probably have someone that can answer to me. And the real simple reason for that is that there is no incentives. And I know some people are working on kind of agent incentives and you know, I'm going to give you more token if you can do that. Or some, someone suggested giving access to better models but like basically they don't have the ultimate incentive to work with you and to perform for you. So that's where a human in the loop is always necessary. Now in terms of what can be done, I'm sorry to say, like we're still testing the limits of it. If you take someone's job and you say like automate everything, I think it's probably the wrong approach. But sometimes when you have specific people that have a better mindset for agentic workflows and they listen to people, people, they actually are able to find an angle in which AI can be useful. And I think developing this mindset inside your organization is the best way to reframe what you do with an AI manner. And I'll give you an example. Working on a flow to help someone sell swimming pools, right? If you were to tell someone like ariance sell more swimming pools, then they probably will gonna just use agent to find more people to target and we're gonna answer the email automatically. But there is actually an answer from an agentic mind which will be like, oh, I can take Google Earth, I can find like properties without swimming po, I can visualize 3D swing pools inside the 3D properties and I can send the vision of like. And this is Completely different. It can be completely automated end to end. And so thinking about it that way and as a person that actually did that like 10x is return on selling swimming pool. But that takes a specific mindset. So I think it's not like what AI cannot do apart from accountability. It's more like how do you think properly about AI workflows to be creative about it. And this is human related.

Speaker A: Awesome. That's a great, great answer. Thank you. I really loved it. Uh, but you know I would be curious because I think I read somewhere in a way acting as kids, they are learning by doing and they hit a limit and they go back and they try a new thing. And I also have a 7 year old that sounds very familiar. I would be super interested to talk to somebody who tried to incentivize agents and maybe ran into then abusing that incentive because they want more or I don't know, whatever, chocolate tokens, whatever the incentive is. I don't know if you know anybody but I think that would be an interesting conversation.

Speaker B: Uh, unfortunately not directly but I read about that a fair bit. One thing we actually trying to do in our is that we start to track our agents based on their performance on specific channels, industry and location. We kind of started to benchmark agents on how much traffic they can deliver, how um, much like deal they can get signed, how much conversion they can get on the page page. Because we started to find that in a world where agent capabilities are commoditized, what's going to win between two people selling similar stuff is the quality of training of the agent and the creativity that's imbued into the skill that these agents are using. Then the question becomes now you can have like thousand skills available online, how do you choose one? And I find that a lot of people answer right now like oh, I use this skill to work well and the gain in productivity is so high that they don't really question that it's the best one. But I think we soon going to get into a land where it's going to get be better to have performance ranked agents. And so the incentive we're going to create is about the performance they're going to have on people's business and growth and bottom line this is an experiment that we started to run about a month ago. I think we'll get good data within three, four months. More than happy to share what we've learned from the agent behavior once they're being tracked. Especially if they start to know they're being tracked.

Speaker A: Okay, yeah, I'LL be happy to do this again to know what you've learned. Awesome. All right, I just have one more question. It's the usual here, and it's about a hack. Do you have a hack? Hack? How to use AI, how to run a company, how to pivot and stay sane, how to not tank in Google after that. Anything that you think other founders would appreciate?

Speaker B: I think there is a, uh, hack that's not a hack. That I think, for me, has changed everything. I think that was about two years ago where I realized that things were changing very fast and that trust was becoming more and more important in the space. It wasn't that much about putting a product in market, because anyone could put a product in market. It was who was recommending this product and who was doing this product, which legitimacy. And so that's where I started to realize that my influence, my personal influence was key. I was an engineer by trade. That was not on social media, that was reactive to the whole LinkedIn space of talking about yourself and stuff. And so I took a very strong step forward by saying, okay, by the end of this year, I want to be that. I want to be visible. I want to leverage my influence. And I'm pretty sure that everyone listening to us will say that they've been impressed with other people being able to create an audience online and that if they haven't done it yet, they will be very keen for it. And so this is where the hype comes, right? Like, so it's still a journey for me. I'm not anywhere near I want to be. But I've been regularly posting every day for maybe a year now, and I see the impact of it. It's amazing. So the hike is this one. I spent a fair bit of a month to wonder why, like, what I would want to be proud of when I was 18, result. And what kind of was the inner drive that was going to motivate me to talk about this goal. And once I found something that was so core to myself that I couldn't wake up every morning and believe in it, then I did the work of aligning why I, uh, should talk about things to achieve that goal. And so I think, for me, the work of authenticity and, yeah, it's okay to put my face out there, came from finding a, uh, deep truth about myself that I was confident about. And then it became very much natural to put myself in front of a camera to talk about the thing I really thought, I think are valuable to the world because at least they're valuable to you and to your goal. I think this hack about working to the bottom of yourself and where you're going and why you're doing it and having an iron confidence into that then helps you get out there and say, like, no, it will be interesting to some people because what you believe is strong, what you believe is very important. I believe, like, marketing is not going anywhere and that people need to get out there and we need innovation to spread faster, faster. And that's why I can talk about it every day. I think that was a great hack that has delivered so much to me so far. And I hope that more founders can find the inner strength to put themselves out there. I'm sure they'll benefit from it.

Speaker A: Yeah, that's an awesome one. Thank you for sharing. And I cannot tell you how much I support this whole trust is important because that's been my bet on this podcast and everything I was doing from day one. Because, yes, it's about also the kind of the, um, specifics of what we're doing at Thought group because you cannot just expect somebody to write you a multi million dollar check if you don't trust them. That's kind of a weird bet. But it's also not just about that. To bring people together that want to relate to you, that want to associate with what you represent is so important. And putting your face out there and not your ghostwriters one is very important. So, yeah, once you find that drive, it really works.

Speaker B: Just to add to that, like, at some point I was trying to take picture of me every day to have a LinkedIn post. At some point I realized, look, let's try to have AI generated picture of me that are very obvious because we're talking about I have every day, like, who is kidding ourselves. And so we started to do that and so now, like, it's worked very well. People love it. I think it's fun. Um, you don't need to be every day there, but I think your thought, your intention needs to be shining through everything you say so that you remain authentic no matter what the layer that you put in between. It's just like it needs to be your impulsion and your desire for the world that shines out there.

Speaker A: Yeah. All right, awesome. Thank you so much. I think that's a great way to finish that episode. It's been great talking to you. Honestly, I think I have a little bit less skepticism towards producing lots and lots of pages on the Internet. I definitely want to try it and see how it works. So, yeah, maybe let's talk after half the record for sure.

Speaker B: And if anyone wants some help with that, our door is open. We'll be happy to share our experience and learn from everyone.

Speaker A: Perfect. The links will be, uh, down there. So thanks again for your time and hope to do it again sometime.

Speaker B: Thank you so much, Anna. Uh, you have a lovely afternoon. Thanks for your time.

Speaker A: Thank you too. Take care.

Speaker B: Bye.

Speaker A: Thanks for listening. SaaS Unbound is brought to you by SaaS Group. We're a long term home for a great B2B SaaS. We buy, keep the team and brand DNA and help with the boring stuff like hiring and finance so founders can truly focus on building great products. If you're a founder who'd like to be featured or explore an acquisition, reach out through the form on our website or email me at annasas Group.