Scaling AI Solutions for Business Impact

This week, I had a fascinating conversation with St John Dalgleish, CEO of Perlon AI, about the reality of building AI-driven solutions that actually deliver results.

We explored how Perlon AI went from consulting to product, the challenges of scaling AI in diverse industries, and what the future holds for businesses using AI.

🌟 Some of my favourite parts of our discussion:

✅ Custom AI at scale: The challenges of tailoring AI models across 15 different industries while staying efficient.
✅ From consulting to SaaS: How Perlon AI transitioned from strategy consulting to a scalable AI product.
✅ AI adoption & scepticism: How businesses are embracing AI—and the fears they still have.
✅ The AI future: Where AI is heading in 2025 and how businesses should prepare.

Business Book Bonus:

https://www.hubspot.com/startups/podcast

https://mistral.vc/the-product-market-fit-show/

https://www.saastr.com/podcasts/

The Road to Conscious Machines by Michael Woolridge: https://amzn.eu/d/07iIzKN

James by Percival Everett: https://amzn.eu/d/gCDVYb2

Transcript

Adam (00:02.468)

Okay, so today I'm here with St John Dalglish, the CEO of Perlon AI, a company whose AI models help clients to leverage advanced data analytics to deliver hyper-personalized outbound communication. St John, welcome to the Fractional CFO show. How you doing?

St John (00:20.62)

Thanks Adam, yeah really well thanks and thanks for having me on.

Adam (00:23.898)

Thanks for being here. Thanks for being here. So, excited for diving into this one. We're going to discuss scaling AI solutions for business impact. So I guess to start with, would you mind giving us a bit of an overview of your career so far? How did you go? I think I saw you were a trader initially to becoming the CEO of Perlon AI.

St John (00:46.234)

Yeah, exactly. I took a bit of a different route into this whole industry. Essentially, the first job that I had out of university was working as a trader for an algorithmic sports gambling company, actually. We would trade markets for UK flat racing using statistical arbitrage. And that was something that I did for about four, four and a half years before deciding that I wanted to really get into startups and start looking at.

Various sort of entry level roles, first of which was at a company called Nested, which was a real estate startup. I joined there as a sort of fairly junior kind of salesperson. Rose sort of threw a few kind of ranks there, building out a business development and sales function for that business. Then switched over into B2B software, working for an American company called Jore. Again, this was a sort of startup, but slightly sort of further along. This was a kind of series B startup when I joined, left sort of after series D once we'd scaled.

that business. And that was really kind of the point at which I started looking into getting a role in the AI space. This was sort of back in 2022. And I joined an NLP startup that specialized in Arabic and what they call low resource language NLP. So that was my first real foray into AI. Spent a while working on some fairly large enterprise projects with the likes of Saudi Aramco, HSBC.

and a few sort of big banks in the Middle East. really kind of the thing that I suppose then prompted me to start Pylon was off the back of that. I saw that there was just this huge gap in the market for advising businesses, both small and large on how to deploy AI effectively for meaningful ROI, essentially. This was kind of, I think, you know, this was like the just about the sort of chat GPT time when that kind of came out. So I've had around a year

working in AI prior to all of the sort of hype and a year at that time was probably a year more than most in terms of their familiarity with the concepts and with LLMs and how they could be utilized. So I joined forces with our CTO, Brent, who's a machine learning engineer who I'd known for several years and we decided to launch Perlon originally as a consultancy, as I said, just advising medium-sized businesses on strategy. Our product itself was born out of that, born out of speaking with a lot of

St John (03:09.958)

potential customers with lot of customers and realizing that there was this huge appetite for wanting to use AI to generate more new business essentially. that's how our platform came about. It was off the back of lot of consulting projects that we've worked on.

Adam (03:28.9)

Excellent, okay, that's great. And obviously, I guess focusing on Perlon and that sort of starting out as a consultancy and then shifting to product, there's obviously a lot of challenges, different challenges as a result of that kind of shift. What were some of the biggest ones that you've come across?

St John (03:49.713)

Yeah, it's interesting. So I've done SaaS before, so I was fairly fortunate in that I was quite familiar with the processes of selling more standardized software. Similarly, my co-founder, Brent, he built and sold a marketing AI copy business prior to founding Pellon with me. That had been used by thousands of different users and we were both kind of fairly familiar, I'd say, with software selling and what it meant to build a product. We were always really kind of working on the consulting gigs that we took on.

with the hope that we would find something that we could productize in that sense. We were just waiting for the right kind of moment. But I suppose that sort of whole selling process is very different. You we were building something suddenly that was needed to be, well, that needed to have application to be used by thousands of users that was built for scale. And that was something that can sometimes be a little tricky to do with a small team. And we very much at the start used our consulting revenue to supplement

the product, to build out the product, to get us to that point where we had enough customers to enable it to be self-sustaining. So yeah, doing that kind of juggling piece at the start was pretty tricky, I guess. probably the biggest challenge is that it really leads to a lack of focus. You don't have the ability to fully focus on your product in the way that you'd like to, because you also need to get some cash in the bank. So I suppose that was probably the biggest challenge at the start.

Adam (05:16.028)

Yeah, interesting. And you mentioned about sort of waiting for the right kind of moment. What was that moment? What was that inflection point where the consulting model gave way to the product focus and what drove that decision?

St John (05:29.649)

Yeah. So for us, it was really just about seeing quite how broad the potential customer set could be for our product. And the great thing about outbound and business development is that pretty much every business in the world needs to do some form of it. And the sort of interest where we put up a really quite a quick, simple, rough and ready landing page for our product. And suddenly, you know, having done a bit of outbound ourselves, having used our product to push our product,

we really started to see a lot of attention from really diverse businesses who wanted to get on board and who wanted to trial the concept. So that was the point at which we suddenly started to think, right, okay, like the market opportunity for this is really big. The other thing obviously about product businesses and software businesses in particular is that the margins are far higher. If you're going to build a product business or if you're to build a software business, you're going to be selling a

a revenue multiple that's probably going to be a lot more juicy than if you're selling a consulting business at the end of the day. So for us that also sort of played into it, you know, which is the biggest opportunity. But we have now very much got to the point where we've spun out both businesses. So Perlon Labs is our separate business, which we have run by a new CEO. We have a CEO called Rohan who came in at the start of this year to run that. And then Perlon AI is very much our kind of our product, our SaaS business. So it was a case of we didn't want to

completely get rid of the consulting business because the speed at which this AI space is moving is such that you really want to have a bit of a finger on the pulse as to what customers and what people are really looking at and what they're requesting and what they're interested in building. So we didn't want to give that up in a way just because of pure intrigue. It's just great to speak to these really diverse businesses that we work with. But yeah, like I say, it needed to be carved off into a separate entity and run by someone who basically wasn't me. I asked if I could focus.

Adam (07:24.058)

Excellent. And yeah, it's interesting that you say about the consultancy acting as that test bed, as that research kind of piece, it makes a lot of sense. And I guess that helps you sort of personalize the solution for different sectors or different clients. But I guess that also comes with challenges. Like how do you manage, you mentioned when we were speaking earlier that you've got tens of clients in tens of different industries. How do you manage the sort of customization?

and not getting overloaded with all of the information from your consulting business and what clients want. How do you manage that customization of AI models for such diverse verticals, diverse clients?

St John (08:05.145)

Yeah, so for our core product, we've got around 45 clients, 15 different industry verticals that utilize our solution. And that responsibility for creating those diverse models sits with our CEO, sorry, our CTO. And frankly, he's gone pretty great over the last 12 months because it is not easy to really manage the different styles of communication that a lot of these customers have and want. know, customers will have their own preferred ways of wanting to communicate.

Sometimes they understand their value prop well, other times they don't. Sometimes their view of what makes good cold outbound doesn't align with what we know makes good cold outbound from looking at all of the data. So that can be quite a tricky process. And customizing for different industries and wildly different businesses is something that has proven to be quite tricky in some instances. But that is also where our system can be set apart from all the others. It's not just a generic GPT prompt wrapper where the messaging

is fairly standardized and easy to spot as AI generated. It's highly nuanced and it is really custom to each business we work with. And that's why we have been able to work with such a broad range of customers, everyone from private jet brokers through to a gold trading syndicate, through to a hedge fund, through to B2B SaaS, through to an interior design business. It's a real mix of customers that we can work with because we put a lot of human effort.

into crafting each of the underlying models that sits behind their access to the platform. But yeah, it's definitely not, it's not easy. And it is definitely spate, spate, well, it's caused a lot of, well, it's taken up a lot of time basically over these last 12 months to really refine what we do for each business. But fundamentally for us, the way that we see it is that outbound sales is something that is generally conducted in a broadly similar way from business to business. To break that down, I mean, the format is,

usually that they'll define the relevant potential buyers of a given product, extract contact details and reach out to them, be that by phone, by email, LinkedIn, it's a fairly similar kind of structure. But it's just that sort of type of communication, the value prop and the target prospects that feed into the system that are fundamentally very different. And that's why you really can't rely on these sort of out of the box models or systems to really communicate that if you're a discerning sales team.

St John (10:21.978)

And you can't go live with Perlon AI in 20 minutes. To create a model properly requires machine learning expertise and sometimes several iterations and several versions. So that's kind of where we have managed to set ourselves apart and where we've managed to pick up some really quite, we call them sort of internally, fairly elite sales teams who really care about the quality of their output and who aren't happy with a lot of the kind of generic crap that's sitting on the market at the moment.

Adam (10:49.576)

Yeah, interesting and you mentioned about spinning out the the product business from the consulting business and having the two side-by-side one of the challenges that I found when companies do that is you know one is the the parent business and the others the child business and there's very much a mindset of this was the original and this is the startup and the the ability to support both those businesses equally give them equal weight How do you manage that as the sort of I guess CEO of the group how it

and I appreciate its early days, but how do you intend to manage that?

St John (11:23.207)

Yeah, so the way that we intend to manage that is for me to step almost entirely away from the consulting business. And that's why we've hired a CEO to specifically run that side of things. My focus is very much now on the product. And so, so is our CTO. Our consulting business, we have some fantastic machine learning engineers, including a new sort of CTO who'll be running that business. So I think from our perspective, like, it is really difficult to make that distinction and it can become very blurred. And for us, it became essential to

make a really sort of clear decision on how we split this up rather than me kind of doing a little bit of each bit of both, because that distraction is ultimately what we felt was going to lead to probably neither business going as far as they could because of your split focus. So yeah, by kind of divvying it up in the way that we have now, this is hopefully what we think is going to be the optimal way of ensuring that both have a really good rate of success going forward.

Adam (12:20.992)

And do you intend to separate it to the extent where you'll have sort of cross charging and separate rate cards for the separate businesses and selling strategies that differ for the different businesses?

St John (12:34.344)

Yeah, totally. We've got different websites for the two businesses now. In terms of the pricing, the pricing for our standard Stas platform is fairly regular. It's the same for each customer, just depending on the sort of scale that they want to do. But the pricing for our consulting projects can be wildly different depending on the time of the commitment. Up into the hundreds of thousands for some of those commitments that we've worked on. So those are very, very custom projects. They're always entirely bespoke.

and they're across a real wide range of different industry verticals. We've worked on everything from computer vision projects through to augmented voice, through to complex bespoke large language models that have been fine tuned for specific purposes. It's a real, real broad range. So yeah, there's absolutely no way we could do any kind of standard rate card pricing for some of that stuff. Also, we've been fortunate to pick up companies in a really wide range of different countries and industries.

which has been fascinating. But again, levels of sort of value perception in different industries and different countries as well is something that kind of varies and we need to bear that in mind with some of our customers. yeah, it's not one size fits all for that business for sure.

Adam (13:48.21)

Okay, great. And you've mentioned you've got quite a lean team. Obviously you're hiring, you've got your new CEO, new CTO, but still a fairly lean team, and working remotely. How has that sort of factored into how you've grown the business, that sort of leanness and working with a remote team?

St John (14:08.702)

Yeah, definitely. We set out initially to try and build a business that was profitable and that made money as soon as it possibly could. We didn't raise any funding at the start. We have raised a small amount of venture funding for our product business since, but our objective was always to try and make money as we were doing things. And as a result, what that meant was that we couldn't add crazy headcount. We had to keep things as lean as possible and we had to rely on assuming that we were never going to raise funding.

so I think that built in some fairly good practices just with the way that we operate. but we do have a sort of distributed tech team. we have several people in the UK. have one person, in Poland at the moment. We have one person in Switzerland, and we're generally looking to build out the technical team in the UK. but with a lean team, you can't do as much as quickly as you'd always like. but you do build this sort of culture of hard work, this culture of profit, culture of revenue generation and like deep.

collaboration is needed when it's a small team. But yeah, for us, we do want to build out the majority of tech team here. We will have to potentially add some headcount in the US. Around 40 % of our customers are in North America. So at some point this year, we are going to put some boots on the ground over there. It's likely to be more on the sort of sales and customer success side though, rather than the tech team, which we'll look to keep here in the UK.

Adam (15:29.32)

Okay, great. And I know building out teams, particularly in your field, can be quite challenging with very competitive market place for top talent. How have you managed that? Any sort of lessons for the audience?

St John (15:43.776)

Yeah, it's really tricky. The way that we've managed it has generally been through network. I'm afraid to say we have enabled through Brent, our CTO, we've enabled access to a pretty good network that he has already. People that he's worked with, people that he's been to university with, and people that we know that we can sort of trust to build our products and to work with us. We've also then got a good group of sort of referral partners and people that we trust on the recruitment side.

who can bring us the right kind of candidates. But it's not easy. yeah, finding the right hires, finding those machine learning engineers that are really going to be the ones to make a difference to your product is a tricky job. mean, to give you an example, we just, we've hired a founding full stack engineer who starts actually on Monday. That search process, I mean, we've literally been looking for like eight months for the right founding full stack engineer. We've had hundreds of interviews and

I suppose we're quite picky with who we bring on, but it's, yeah, that's been an eye opener just as to how difficult it can be to really find the right person. But it also just talks to the fact that you don't want to compromise on quality. You want to really make sure you get those right hires. They've got to be the right fit because when there's only a few people in the business, firstly, you've got to like the person. Secondly, they've got to be someone who is prepared to pull out all the stops. They're going to be someone who's going to work just as hard as the founders to get this product off the ground and to get everything done. So.

It's yeah, it's not easy to find that person, I think.

Adam (17:15.944)

I mean, how do you do that? How do you keep that person motivated? You know, you've got a small team, some people remote, you're working incredibly hard because it's a startup and always iterating. How do you keep them efficient? How do you keep them motivated? What are the secrets?

St John (17:31.235)

Yeah, so in terms of the team that we've hired up until this point, they've all been technical. So those technical team members have sat with Brent, our CTO. And I think from his perspective, it's really just been about ensuring that we're working on truly interesting things throughout the day and building something where we can see like really tangible benefit of every sort of line of code that's written and getting that buy-in I think has been really key for him.

But in terms of getting the best out of other people, I'd have to ask Brett exactly what tactics he's using there because he's actually managing the technical team on a daily basis.

Adam (18:06.46)

Great. Just shifting gears slightly, you're obviously working in a very fast moving field and a lot of businesses don't fully understand AI yet. How have you approached educating customers on the potential of AI when it's something that's so fast moving?

St John (18:27.203)

Yeah, it's really interesting. And we've worked on consulting projects for companies at every stage of that sort of AI transformation cycle and every stage of understanding when it comes to what's possible. first ever customer for our consulting business, it was literally a case of we went to their offices in London. We sat in the room and the question they asked was how, how can we use AI and what is AI to our business? And it was like, right, okay, let's break this down. Let's look at every single element of what it is that you do.

and see how it can be applied. Other customers will come to us with a more advanced understanding of what's possible and they'll have an idea of what they want to do. But we generally start with something of an AI audit, let's say, to fully understand the business's processes. And we work from there on really determining where the opportunities could lie. But the great thing about, well, the great thing about ChatGPT is that you've just had this explosion in the interest in this space.

So you really, going, you're not, you know, I'm not going into these companies and saying, you know, we're going to show you how to optimize your CRM or optimize your Excel sheets or some other boring software applications that they're forced into using. Like we're going in there and saying, right, we want to look at how we can apply AI to the business. And the amazing thing, like said, is that people are so into this. People love AI. They love chat GPT and they're actually really engaged with seeing how this can benefit them.

in the long term and in the short term. So that's been really great and that's been kind of something that I suppose I hadn't really expected. It's just that level of enthusiasm that you get from so many people at every level within a business. So I think that sort of buy-in and the willingness from teams to adopt AI has been amazing. So that's something that we've really benefited from, think.

Adam (20:12.424)

And obviously you've got that excitement and enthusiasm as you say, but I guess there may also be some misconceptions or even fears that clients have about AI. Have you come across any and how have you addressed them if so?

St John (20:26.63)

Yeah, it's interesting. We have come across that a bit of a bit of pushback with certain teams, even with our product, actually certain teams who let's say they're a big, a big chunk of their daily workload is doing outbound prospecting, right? Over email, over LinkedIn. you know, even the concept of the fact that that can be, you know, automated in a really effective way. can sometimes you notice it be a little bit worrying to certain people. It's often not very explicit in the way they communicate that it's just kind of, can, you can sort of slightly tell.

with the way that they react. It's generally the people who are fundamentally doing that work themselves rather than the managerial roles where you kind of see that. I mean, to be completely honest, there have been a few projects that we've worked on. And I won't go into the specifics, but there are a few projects that we've worked on for the consulting business where the express mandate has been to reduce full-time headcount from using an applied LLM or from using a particular tool that we're able to build.

So you can kind of see how in some instances that concern that people might have is definitely warranted. But a lot of the teams that we work with on the flip side will want to use applied AI so they can hire more people, so they can do even more business. And so those people's time can be diverted to exercises that are way more reflective of actual human skill and their ability to close deals or whatever that might be. So I suppose you've got kind of two sides to the coin really. You've got those teams who will...

for whatever reason and for mandates that they have want to cut headcount. And you've also got teams who'll see it as a way of just, you know, fueling even further growth. So, so yeah, you're always going to get some sort of pushback, I guess, to an extent with some of these concepts.

Adam (22:06.354)

Yeah, yeah, of course. And I guess following on from that, it's obviously evolving incredibly rapidly, the space. And so how do you stay ahead of that? Both yourselves staying ahead of the industry changes and then staying ahead of the competition. You mentioned at the outset, how you differentiate your products slightly, but it'd be interesting to understand your kind of R &D approach and your thinking around how to keep ahead of the game.

St John (22:31.519)

Yeah, sure. So I can sort of talk to that on both levels. So for our product business, our CTO is constantly testing the newest models available on the market, looking for incremental improvements to the quality of the outputs that they can fundamentally generate. The issue I suppose that we've had with LLMs and our product is that these LLMs, they're benchmarks on maths, they're benchmarks on reasoning, they're benchmarks on their ability to code. It's all fairly...

quantifiable stuff that sort of forms the, I suppose the opinion as to which of these LMS is the best for whatever purpose. The thing that they're not benchmarked on and the thing that's pretty qualitative is how human does your output sound? And you've probably been on the receiving end of some pretty bad AI generated emails. I just received one of the worst ones I've ever had this morning actually. And that ability to send something or to create a message that sounds very human is not something that's easy to benchmark in any meaningful way.

And this is the thing that we've noticed with a lot of these models and the updates that keep coming out is that whilst we've managed to manipulate them in such a way that we get pretty good human sounding emails, like the step changes just aren't coming. So for us, in terms of that sort of quality and in terms of like the space evolving, we are actually fine tuning our own large language model using an RLHF technique that will help us to really get to that sort of next stage of quality when it comes to output.

In terms of the sort of space that's evolving around us, I mean, that means we'll have our own model version that sort of sits behind our main platform. But we are fundamentally an AI positive company in the sense that any improvements to the sort of, you know, the foundation models to the traditional models that are available in the market, they're always going to be good for us because they'll just help to make our product better. If we decide that our own fine tuned model isn't as good as, you know, the latest GPT, then we'll just switch over to using that. So.

We're quite lucky in that sense in that all improvements in the space will help us deliver better results for our customers. But we have sort of, we have really kind of determined that like these algorithms just aren't really fit for purpose in many ways in the way that they currently sort of are formed. So that's, that's sort of one side of things on the, consulting business. We have, you know, our first hire on that business is our head of R and D. So that is to ensure that every project that we work on is using the latest models.

St John (24:51.948)

is using the latest techniques. And that really kind of helps us to stay ahead of the curve there. I think you've got to hire people who are heavily invested in AI, who are heavily invested in like what these latest techniques are and what the most, I suppose, interesting things that are developing in the theory of this space are. So that's really sort of been quite important as well. It's just making sure that it's a team that are very keen on understanding what those changes are and when they come about.

Adam (25:22.17)

Interesting, yeah. And I guess a difficult question now, but where do you see it going? Where do you see the AI landscape going over the next year, two years, three years? I'm sure you do some sort of planning, but obviously it's such a new industry. How do you think about these things?

St John (25:37.899)

Yeah, it's a really, really difficult question to answer that because the pace of change is just insane. And I think what you're going to start to see is obviously we have an agentic system in our main sort of core platform. And I think that this 2025 is certainly going to be the year in which these agents really start to become a lot more prevalent agents for a wide range of different reasons. Essentially our agent, our core agent that we call it internally is

It's for sales. It's for outbound sales. That's, that's, that's effectively a sales agent. Um, but the, development of those agents, those autonomous agents that can be trained, that can be told about your business and that can then go and do the work themselves is probably what we're going to see a lot of this year. The experiments were starting in 2024. Some of the early adopters were picking these things up. Some of the rough and ready products were being released, you know, including our own. Um, and I think 2025 is really going to be the, the acceleration of that kind of agentic framework. Um, beyond 2025, um, it's.

It's really, really difficult to put pen to paper and to sign off on any kind of predictions as to where this whole space goes. A lot of what's happening is pretty scary. A lot of what's happening is really exciting. And I think this is the thing for us is we've just got to stay as far ahead of the curve or on the curve as possible and just speaking to as many customers and potential customers as we can about what they want, what they see as beneficial and how we can sort of use the supply they are in the best way for them basically.

Adam (27:06.802)

Yeah, challenging for sure. And I guess segue on that is like your journey as a founder and your journey has been through different technologies and through different roles to get to where you are today. And what have you found to be the most challenging part of that journey to get to where you are? And then how has that shaped your style, your leadership style, your company culture?

St John (27:09.251)

Yeah.

St John (27:33.604)

Yeah. So I've worked in startups for a little while prior to starting Polar. So I was kind of familiar with some of the ups and downs that go with startup life. a of climatized that and to an extent, but I think when you have a small team and it's your business, what you really kind of start to realize is just how little time you have off and how little time you have to really be able to kind of switch off. And I think in terms of the challenges that's probably been

the biggest one, like everything falls on you. I personally, for example, you know, up until this point, I dealt with all sales, every deals come through me. Every bit of customer success has come through me. You know, the phone is never off. Like the customer service support line is me. It's my mobile basically. So I think that has been the biggest challenge is just really trying to navigate the fact that you can't ever really switch off.

in a way that you can when you're employed in a role. I used to remember a couple of summers ago when you just go and take a week off, you check your emails once a day or so, but who cares, you sit on the beach kind of thing. That life doesn't happen anymore. So I suppose that's been a bit of shock to me, to the system. But yeah, all part of it. think I'm probably just about used to it now after a year and a half.

Adam (28:54.952)

Yeah, no, absolutely. can definitely relate to that. I can definitely relate to that. And I guess if you, you having started your own thing now, it's obviously very different from being an employee, as you say. You know, a lot of the audience to this podcast are people who are starting out on second careers or thinking about it. If you could give one piece of advice to someone launching a B2B business in 2025 this year, what would that be?

St John (28:58.372)

Yeah.

St John (29:22.023)

Yeah, it's interesting. think that the stock advice that I see a lot of VCs giving at the moment is find like a really small niche and build for it. That's sort of what our sort of original Perlon products was. We actually started building our outbound product for private aviation, for private jet brokers. Those were our first eight customers. So that kind of quickly then grew into something that was very much a horizontal product that was used by a lot of different industries.

But I think that sort of advice, yeah, it can be, it's interesting. And I think if I would start another B2B business, I definitely target an industry that was quite niche and that was probably a little bit old school in the processes that it had. And that's been sort of slower to adopt AI in any form. I mean, to give you an example, like we've recently worked on a really interesting and highly complex custom LLM for one of the largest infrastructure construction firms in the Middle East.

And again, like this is a sort of business that is entirely reliant on physical, you know, goods, physical building, like actual engineering of the, of the old school kind of kind, and not the kind that we're familiar with in-house. And again, there is a lot of, there's a lot of room to optimize the processes that they go through and to utilize AI throughout their business. And again, I think it's just trying to look for some of those niches, some of those sort of slightly more

forgotten kind of old school industries potentially where you can really look to improve things with technology. That is probably the direction I would go in if I wanted to start another business. And yeah, think on a sort of personal note, my advice to anyone who would be starting a business now would be to prepare for the sort of catastrophic daily highs and lows.

You know, when you're employed, you don't notice these things as much, like when it's essential at this level, when it's your business, um, to be as calm, um, and, and, and I suppose level headed as possible when you're going through, you know, really bad days and really good days, like, this is, this, this is a weekly occurrence for me still, you know, you'll have something go disaster. So you're on one day and then the next day something great will happen and you've just got to be able to manage that. So I think on a personal note, I definitely just try and get people ready for, for that. Cause like it definitely will come.

Adam (31:39.666)

Yeah, absolutely. Definitely helps you grow as an individual, that's for sure. And that resilience that you require. Exactly right. Excellent. So just changing tack and moving on to what I call our business book bonus section. This is where we ask our to provide us with a recommendation for the audience of a business book or any business content that's really helped you.

St John (31:46.09)

Yeah.

Adam (32:03.408)

during your business career and that you would want to recommend? So what business book or other content would you like to recommend?

St John (32:10.103)

Yeah, so I'm quite a big fan of podcasts. So there's three that I've been listening to a lot over the last year or so. The first of which is going to sound kind of lame actually. It's a HubSpot podcast. Actually all three of these podcasts that I listened to are kind of like branded company podcasts, but they're pretty good. The HubSpot podcast is called The Science of Scaling. So that's really good as a, again, they get a lot of founders on there who are talking about their journeys of building businesses.

they've got some pretty good people that they have on there from, from open AI and from, from a various other companies. So that's, that's a good one. Just really kind of exposing like what it takes to scale a company again, fairly sort of digestible format. They're like 30, 30 minutes or so. the other two that I really like again, like this is kind of. And tech focused, but the product market fit show, which is produced by Mistral, which is another VC. they have some really good guests to explore. yeah, again.

It's achieving product market fit. speak to companies at various different levels. So small, large, it's a real interesting mix of guests that they get on. And similarly, there's the Sasta podcast, which is again, very much focused on Sass businesses, various ones that have been venture backed, various ones that have been bootstrapped. Again, very specific to the industry that I work in. And again, I might sound like a real kind of tech bro when I say this, but like I often do listen to those.

for guys on the all in podcasts. find that some of the topics that they cover are quite interesting. They're pretty odious characters in general, I think, but like what they talk about can sometimes be pretty good and pretty relevant. So I think that's kind of where I go for sort of podcast content. When it comes to sort of books, if you're interested in AI and getting a bit of an understanding as to sort of how its progression has mapped over the last few decades, there's a book called The Road to Conscious Machines by Michael Woolridge.

which is a really good intro book. He's a computer science professor at Oxford, but he makes this book fairly digestible for non-technical readers. And it's a really good timeline of what's happened and where the sort of future could potentially go. So I'd recommend that to anyone interested. And then fundamentally, other than that, be completely honest, I've found recently that switching off about outside work and reading fiction has been a really enjoyable experience for me. I'm kind of finding a bit of solace and detaching from all the AI startup.

St John (34:29.965)

business BS to be honest and just like getting back to actually losing myself in a bit of fiction. I just finished a book last night called James by Percival Everett which was the story of Huckleberry Finn told from the perspective of a slave. I'm kind of just trying to detach a little bit from the business stuff at times.

Adam (34:46.514)

Yeah, you've got to do that sometimes. Absolutely. Great. Okay. There's some great recommendations there. So we'll, we'll put links to all of those in the show notes. Thank you very much for that. And I guess before we wrap up, there anything that we haven't covered that you'd like to mention or where can people find you if not?

St John (35:02.843)

Yeah, that was a really good conversation, Adam. So thanks for that. I appreciate it. And yeah, I'm on LinkedIn. You can find me, Singen Dahl-Gleisch, and our companies are, yeah, PerlonLabs.com and PerlonAI.com. So Perlon Labs is for our consulting and PerlonAI is for our outbound sales product. So yeah, happy to chat with anyone. I'm still very much involved with all of the sales processes at both businesses. So yeah, be happy to chat with anyone who wants to talk about either of those two concepts.

Adam (35:31.912)

Okay, great. Well, thank you very much. And thanks for joining me on the Fractional CFO Show, Sinjin. Really appreciate your insights, your perspective and your time. Thank you.

St John (35:40.228)

Thanks Adam.