.png)
SaaS Stories
SaaS Stories is my not-so-secret quest to learn what it truly takes to succeed in the world of SaaS—and I’m inviting you along for the ride! I have the pleasure of sitting down with brilliant minds and industry trailblazers to explore their journeys, uncovering the secrets behind their growth, the gaps they spotted in the market, and what really drives them.
It’s not all smooth sailing—there are challenges, unexpected turns, and moments of reflection where they share what they’d love to change about their journey. Think of it as a candid, insider’s look into the world of SaaS, with just the right amount of curiosity, empathy, and wit.
Join me as I dive deep, selfishly soak up all the insights, and hopefully share a little inspiration with you along the way—one SaaS story at a time.
SaaS Stories
How SSW turned AI into ½ their pipeline — Ulysses Maclaren, COO of SSW
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Most organisations are giving their teams AI tools and hoping for the best. The ones actually pulling ahead are doing something completely different and it starts long before anyone opens Claude.
In this episode of SaaS Stories, Joana sits down with Ulysses Maclaren, COO of SSW, one of Australia's leading software consultancies, to unpack what 20 years of technology transformation actually looks like from the inside. From .NET development shop to AI consultancy, Uly has lived through every major tech wave and knows exactly what separates the organisations that adapt from the ones that get left behind.
This conversation gets specific. We cover:
→ Why giving everyone a Claude licence without a strategy is a fast way to waste money
→ The enterprise vs BYO AI adoption models and the trade-offs nobody talks about
→ Why exec buy-in matters more than any tool purchase you'll ever make
→ How developers are shifting from writing code to managing "AI clankers" — and why some of them hate it
→ The inside story of YakShaver, the internal tool SSW built to turn a screen recording into a perfectly formatted backlog item, and what the go-to-market taught them about building for yourself vs building for a market
→ Why "selling to everyone" is a fast way to sell to no one and how to find the niche that actually wins
→ What to look for when hiring for an AI-native team in 2026
🎙️ SaaS Stories is brought to you by Hat Media — B2B marketing specialists for the SaaS world. Subscribe and catch every episode at hatmedia.com.au/SaaS-Stories.
Welcome And The Stealth Interview
SPEAKER_02Welcome everyone to another episode of SAS Stories. I'm joined today by Ulysses McLaren, CEO of SSW. Welcome.
SPEAKER_01Thank you very much.
SPEAKER_02Now you've got an interesting story. You've been at SSW for 20 years, which is pretty rare in the world of tech and SaaS. Take us back to the beginning. Tell me what is the origins to now COO of an amazing company that's one of the leading software consultants in Australia.
SPEAKER_01Right. Okay. Well, I I joined, I was working at Foxtel as an encryption engineer out of uni. And a friend of mine was doing work experience at SSW. And he's like, this is cool. You get to work as a software consultancy, you get to work with all different companies, all different industries. It's much more interesting because he'd done work experience with me at uni at Foxtel as well. So he's like, you should come over and check this out. Um and so somehow he convinced me one day to come and just do a ride along, like, come and check it out. Let's see what it's like. And I believed him that I was just going to go check it out. And I didn't realize it was a stealth job interview. Um, and I turned up, I used to have long hair, and I uh was teaching uh dancing at the time, salta dancing, and I was doing a performance and I had it in cornrows. So I'm a what, a 23-year-old uh six foot four English white person with cornrows who turns up at this software consultancy, and somehow I ended up with a job offer at the end of the day. Um, Adam somehow saw past all the weirdness and liked it. Uh, and I ended up taking a big pay cut and moving over to a software consultancy, but it's been gangbusters since then because it's yeah, 20 years, you're right, it's a long time to be at one company. But software consultancy is just the best industry to be in because technology changes constantly and never goes stale. You're working with different industries, different problems, different puzzles, different people, different personalities. Um, it's it's an endlessly fun puzzle to solve.
Why Consultancy Never Gets Stale
SPEAKER_02And so technology does always change. Um, what were like the most common problems that you saw throughout the 20 years? And how are they different from now with AI playing a role into technology and working?
SPEAKER_01Well, so I started off as an account manager uh and then I became general manager and then COO. Um, but all the way through I've been doing account management. Uh, and I tend to get very excited by any technology that allows executives more power, weirdly enough, working in a software consultancy, that needs as little help from developers as possible. So the big two things that excited me in my career have been uh Power BI and business intelligence in general, and those sort of low-code ways of self-serving yourself with what you need. So as an executive, I need to know what the profit figures were for you know this subset. And instead of sending that off to a developer and waiting a month, right? Or a week even, like I can I can answer my own questions. Uh and AI is like that on steroids, because you can you can self-serve now, you know, with vibe coding and everything, you can even build your own software. Um danger, danger, but still cool. Uh so yeah, look, AI is is the next big wave. And when uh in in November of 2022, when GPT 3.5 came out and we all got very excited, uh, I was already excited. Right. And so when this happened, uh I talked to Adam and we decided to pivot the company pretty hard from just a pure enterprise software company to an AI consultancy.
Pivoting Hard Into AI Consulting
SPEAKER_01Uh and that was a big move. It required a lot of rebranding. Uh, but I think it's the right move, and it's definitely where everything's going. And and again, it's it just allows so much more individual power for anyone in the business. You can you can stray outside your lane a lot more than you used to be able to now.
SPEAKER_02Yeah. It definitely, I mean, it definitely is the future. And depending on who I talk to, I either feel really dumb in AI or quite advanced. And um, I'm curious, like what kind of companies come to you and what questions around AI do they have? Like, what is the biggest problem they have? Is it too many tools? Which one do we use? Or how do we systemize all these things?
SPEAKER_01It's it's changing a lot over time. Um, right now, the biggest question we're getting is hey, we want to give our whole company Claude and they don't know how to use it. Can you please tell all our people how to use Claude?
SPEAKER_02Guilty.
SPEAKER_01Yeah. So Clawed, Clawed code, cowork. Um, just that one tool is now so complicated and so wide that it requires a whole training course just to sort of you don't just give them uh you know a ChatGPT license that it used to be, and then you know, oh, it's a chatbot, you'll figure it out. Like now you you you can be a very good power user with these things, or you can be a surface user. So that that's it's a it's not much income for us to get that you know training work, but uh it's often a foot in the door because they start using the tools and they start going, they they start realizing where their barriers are, but they also get more ambitious because they're like, oh, hang on, I could almost do this and I can see that it's possible. Can you come back in and help us build that? So it's funny because we've gone from uh most of our jobs used to be quite big jobs, like you know, three, six month jobs, uh building a piece of software or or modernizing a piece of software or moving something to the cloud or something like that. Now we're dealing with, you know, two weeks, one month jobs where you get a lot done really, really quickly and then you move on to the next thing. So there's a lot of RPA robotic uh process automation.
unknownYeah.
SPEAKER_01Um, with just basically we have a workflow, uh, we tried to turn it into a skill, we got stuck, help. Um so really little simple jobs, but lots of them.
SPEAKER_02And
Getting AI Adoption To Stick
SPEAKER_02how do you find companies help their employees adopt AI? Because I feel like a lot of them, you know, the CEO sort of saying, everyone needs to use AI, but then there's no actual education, there's no systems, even security, there's some issues there. And um, you know, either people are too scared to use it and treat it like ChatGPT and to sort of ask it basic things like write this email for me.
SPEAKER_00Yeah.
SPEAKER_02Or they go on a bit of a connection frenzy and um that causes some issues with, you know, um security.
SPEAKER_01There's there's very wide-ranging ways to go about it. So you can you can go whole hog and buy team or enterprise licenses and give the seats to your employees. Um obviously the good thing is there everyone's enabled. Uh, but the downside is often you have adoption dramas because people, even though they have access to it, there's a lot of um status quo bias, right? I I've done, especially for people who've been at a company for 20 years, I've done this for 20 years, don't tell me how to do my job. Right. So um that that's uh uh a challenge to get past. And the best way to get past that is to uh start from the top. You've got to get your executive suite uh buying in. You've got to get the the owner of the business, the founder, the the COO. I I feel like my position is the best position to influence the business because one, you're coming in from a leadership position on multiple teams, you're able to see areas uh just from uh from a high level zoomed out where they're doing things slowly, where they're stuck on things that don't need to be, where they're doing things repeatedly that that they don't need to, uh, and then you can help educate. So educating your C-suite first, I think is a big deal. Uh the other way to do it is instead of uh supplying it for the business, you've got your BYO AI, right? So everyone bringing their own AI, right? And then you maybe subsidize what they're doing. So uh you could say, you know, for instance, hey Bob, if you want to use AI, uh you can bring whichever AI you want. We're not dictating you don't have to use anthropic or open AI or whatever, whatever you vibe with, go with it. Uh, and we'll pay for all or some of that. Um keeping the employees with some skin in the game, they're paying a small sliver of what they're using, uh, does mean that you don't end up with a problem of uh a low adoption because if they're not going to use it, they're not gonna pay anything for it. Uh and people do get a lot out of it in their own lives as well. So often, especially in a software consultancy with developers who would be doing this on the weekend even if we weren't paying them, uh, you know, it there's it's a good model for us.
unknownYeah.
SPEAKER_02Yeah. I definitely want to talk to you about people management and how you've encouraged people to use it and all that, but we'll get into that later. I just want to close off on AI just for a minute. Um, because SSW essentially transformed from a.NET company into now, I think it's something like 19% of revenue. It's in the product and the pipeline.
SPEAKER_01So um how I'm scared that you know that.
SPEAKER_02I've done my research. Um curious, what has been essential for the survival of a 20-year-old company? I mean, obviously you've adopted new technologies and that's been important. Is there anything else?
SPEAKER_01Um Adam Kogan is our chief architect, and he is a Microsoft regional director. And I think that has helped us a lot with uh navigating the channels of changing technology over time. Because, you know, for for ages, for the longest time, you know, the which JavaScript framework should we use was the biggest question. Everyone was competing over React and Angular and Vue and whatever other ones. Uh, and there'd be whole sort of parts of the business that are saying, no, this is definitely better. We've got an office in China. The Chinese audience is has their own preferences, uh, Chinese companies have their own preferences, Western companies have different preferences. So trying to standardize across that kind of stuff was really hard. And having uh Adam as a regional director, which means he's one of, you know, a select group of regional directors who can all ask each other these sorts of questions. Hey, are you guys using mainly React? Or no, no, we're sticking with Angular or whatever, um, really helped us like get through that channel. And I think everyone needs to have an external source of um a brainstrust, you know, because you can get very stuck in a business just doing what you all do uh and ending up in an echo chamber and not seeing what's new and what's out there and what's changing. So but also you can get overwhelmed by what's changing, especially now with AI, right? Because it changes weekly.
SPEAKER_00Yeah.
SPEAKER_01Um, so you need you need to be having people who are going out pinging other people and figuring out what the best thing should be, uh, and then discussing amongst yourselves and standardizing.
SPEAKER_02Yeah,
Cost Of Change Drops Fast
SPEAKER_02it sounds like you kind of need two lots of people, right? The ones that are excited by change and innovation and go out and learn about it and how do we move the company forward. But you also need the ones that sort of provide that stability and keep everyone focused. So like don't go chasing the shiny objects all the time because actually the revenue's coming from here and we need to service these clients. Um, do you does does the team sort of work that way?
SPEAKER_01Yeah, so that's the the visionary and integrator model, right? So you've got your vision, like most most great companies have a visionary, hopefully CEO and an integrator COO who uh the visionaries has ADHD and goes off and gets excited about all the things, and the integrator goes, All right, all right, hang on, let me get that's a good idea, let's go with that and let's let's spread that out and make that happen. Um and you need that, as you say, at different scales as well. Um I would say, and I'm gonna sound self-serving here, but I would say that now it's easier to jump ship. The the cost of change has reduced. There is more change because things are changing really rapidly, but also you can get AI to do a lot of the uh plumbing work um that was the big thing that would slow you down before for making a big change. Right. So uh a simple slightly technical example. Um I vibe coded up an MCP server to hook up to uh Meetup, right? So it's just uh like an alternative to an API so that I could my AI could talk to Meetup, right? Because we post events and I wanted our events to be up on these things without my admin person having to spend half a day putting in all these different places. Uh there was a new spec, a new version of the um MCP spec came out. Uh normally that would have been a whole thing. Developers would have to spend a sprint working on it. I literally uh we call it we call AI clankers. So I told my clanker to just I said, here's the new spec, can you update it, please? And within five minutes, it was done. So the cost of change has reduced significantly, which means you can afford to chase the shiny ball a bit more nowadays, which is really fun. So I I would say you do need to still be careful. You can't always be just looking at what's new and not actually working out where the where the ROI comes from. But if you see something cool and you think there could be ROI, it's cheaper to find out nowadays than it used to be.
Developers Become Managers Of AI
unknownYeah.
SPEAKER_02And you've got a lot of developers in your companies. Um a lot of companies, not gonna name names. They've been in the news, but um, they've had so actually because of AI, at least that's what they've claimed, they've made a lot of people redundant thinking AI is gonna solve the problems. Three months later, they're having to rehire everyone. In SSW, are you having to retrain what the developers do? Or are they just naturally shifting towards new tasks? Do you find that the type of work that they're doing is changing?
SPEAKER_01Software developers, i if you're a software developer, if you go into a career as a software developer, you know you're signing up for a lifetime of learning, right? You you're not gonna go to uni like back when I went to uni and learned, you know, Java or I learned Flash Action Script. There you go. Right, right. Or uh OGC or or assembly language or whatever. Like whatever you learn at uni is not what you're gonna use in your career. And what you use in the beginning of your career is not gonna be what you use in the middle of your career, and so on, right? So the the frameworks, we're used to the frameworks changing. Um Having said that, it has still sometimes been a slog to get people to not just change language and just because they they see the value of that, they they that's the kind of change they're used to, but change their entire process to go from like obsessing over the syntax of of what they're writing to actually zooming out from that, abstracting that up one layer, and you know, commanding a horde of clankers to go and do your bidding for you, and then just look just checking their work. It it's more like a manager's role than it is like a software developer's role. And that's that different in the style of work, some people don't like at all, right? They they really are happy when they're in the zone coding and and they don't want to be told that that's not an economically useful thing to do anymore. So that's hard.
SPEAKER_02I don't know how I feel about some of the work that I love. Like lately, I feel like all I'm doing is briefing different AIs to do different things, but at the same time, I've never been more productive than I am right now. So um it's just a mindset shift in a way.
SPEAKER_01Yeah, look, you've got to have the right, the right mindset, right? Uh first off, you can't it it's very easy to cheat, right? We're talking to a lot of companies in education, and most people in education are more scared of AI than they are excited by it because students cheat. Students, like in school, are low motivated learners. They're there, they're learning because they have to, and they do their homework because if they don't do their homework, then they get told, hey, you didn't do your homework. They don't really care about the outcome, right? So they were they are the prime candidate for someone who will literally put the homework question into Claude and then copy and paste the answer into the thing and they learn nothing, right? Um however, AI is also the greatest enabler for education that has ever existed. Like I didn't do any AI machine learning, anything like that at university. I'm now, you know, uh like a leading expert in it, and I've learned mostly from just talking to AI about AI. Um, you know, if you get a new paper that comes out, you put it into Notebook LM and it outputs a podcast for you, which is my easy way of learning. I love listening to podcasts and just having people chat back and forth about a topic. And then I can interrogate it and I can go, okay, can I just check my understanding here? Did you say that blah, blah, blah? And I'll say, not quite. Actually, I think you're thinking of this, it's actually like that. Like it's like having a we call it a digital Socrates, like a personal one-on-one tutor, right? So this paradigm enables curious people, it enables um ROI-focused people to look at something, understand it well, and extract the useful parts from it. So if you use it, if you use AI intelligently, you are, you know, 10x, someone who uses it, or who doesn't use it at all. And the people who use it because they're told to use it, but they're not really understanding how aren't much better than the people who don't use it at all because they're giving mid-answers, right? They're giving the standard, you know, obviously written by ChatGPT, overly verbose, full of M-dashes. Like you can spot it a mile off if people are using AI poorly. Um, yeah, so it's it makes a big difference.
SPEAKER_02Yeah, I agree. I agree. You can either use it to enrich your brain and education, or you can completely make your brain redundant. Um, I was really proud recently I used it to fix my broken dishwasher, which work, and make a killer beef nachos recipe.
SPEAKER_01So I was like, yeah, I look, I honestly, I think, I think home use cases are the best way for people to start thinking, oh, I could use this at work. Like simple stuff like on my phone, I've got a uh it's an iPhone and there's like an action button. So my action button, instead of being used to turn it into silent or whatever, I press that and it opens up ChatGPT in voice mode. Right. And then you can turn the video on with that or take photos with that as well and just chat your way through. So I'm doing simple stuff like the laundry, and I find my some piece of my wife's clothing that looks a bit scary. Oh my god, I'm gonna put this in the machine, it's gonna die. It's got some kind of hieroglyphics on the label that explain what you can and can't do with it. The hieroglyphics, I don't know who invented this, but I don't understand it. So so I just get ChatGPD to talk me through. Like, what can I do with this? Well, can I put this in with this? No, don't do that. What are you trying to kill it? You know, it it's fantastic. It's again, it's like a little little helper. Um But the the I the idea of cognitive offloading is a big deal. Um and I've always liked that. So I've always I think that's that's a philosophical thing. So I am I'm not proud about how smart I am. Like I don't think I have to remember stuff in order to be smart, right? So pre-AI, I already use lists for everything. There's there's a book called Getting Things Done, um, which is fantastic. And it it talks about basically having systems for everything. Like, don't store anything in your head. Your head should be for in the moment processing, not for, oh, don't forget you've got to juggle this thing and that thing, and don't forget to do also this, and you need to get back to this person and all that kind of stuff. So if you've got systems for everything, then it it allows more of your RAM in your head to focus on what you're doing in front of you. And AI is another way of cognitively offloading things. So you can focus on bigger things. You can you can realize, all right, AI is gonna, if I've given it enough context, then it's working with me on a problem. I have a pair, a pair brain to think on a topic with.
SPEAKER_02That's a really interesting way of putting it. I was just thinking as you were saying that, it's probably easier for a male brain to do that than a female brain, because we've got 20 different tabs opened all at once.
SPEAKER_01Yeah, maybe. Yeah, maybe I'm going, I'm going deep male. I don't know.
SPEAKER_02Um,
Cognitive Offloading And Better Learning
SPEAKER_02shifting focus to Yakshava. This is your
Yakshaver Turns Video Into Tasks
SPEAKER_02baby that you built within SSW. I'm curious how what was the idea behind it? How did you launch it? And also um for all the marketers and sales leaders listening to this right now, what were some of the scaling um tips and advantages you can share with us as well?
SPEAKER_01All right. I'll answer one of those questions and you can remind me of the others because they'll forget. So, first off, uh SSW is a consulting company, but we also sell a couple of products, right? But it primarily consulting. Um, so all the products that we build, we build because we want them. So for instance, we have you know a timeshooting system, an ERP, a CRM, like uh an e-learning system, an onboarding system, all the things that we couldn't find off the shelf, and we've got devs to spare. So we're like, we'll just build the perfect one for us.
SPEAKER_00Do it ourselves. Exactly. Yeah.
SPEAKER_01So Yakshaver at the beginning was a tool for us. And we've talked a bit about Adam Kogan, who's the owner of the business. Adam, if you're watching, I'm gonna say things about you now. Uh Adam is incapable of walking past a bug. Right? So this was this was the problem that we had in the business. If you have a meeting with Adam, uh, you had no idea how long it would take because it depends how many uh, you know, sparks pop up, like sparkly things pop up in the in the time.
SPEAKER_02The visionary.
SPEAKER_01The visionary, exactly. Yeah. So so a lot of any meeting with Adam is calling other people into the meeting to talk about something you just noticed and getting that into the backlog for whichever product or or team needs to work on that thing. Uh and it was uh challenging, right? Because we couldn't schedule our time. If you had a meeting with Adam, it was it was a write-off. Uh we'd never get through a board meeting, right? Um, so we decided to fix this problem with software. So what we did is we wanted the quickest possible way for someone to notice, let's say, a bug in a piece of software in a report, whatever, and get that to the team that needs to fix that bug in a way that they can understand that they have enough context. So um, you know, originally it was typing it, right? But there's a fairly low bandwidth way of communicating. It takes a while to properly write up a user story. Uh, maybe you put a screenshot in, you probably call the people in so they understand it before you send it, all that kind of stuff. So it all takes ages. Uh I used to be obsessed with Siri. I like the idea of voice. Yeah. Um, but what's, you know, they say a picture says a thousand words. So a video is a voice and a picture all in one. It's very, very high bandwidth, right? So before we get into what is it, uh, brain computer interfaces, that's the best we got. So what Yakshaver does is you just press a keyboard shortcut on your keyboard, it records your screen and your voice. You walk through whatever the drama is or or the idea you have, uh, you press stop, and then the AI does all the Abbin for you. So it goes, right, which project were they talking about? What uh backlog system is that using? Is it in GitHub, Azure DevOps, ZenDesk, just emails, whatever? Uh, how do they like to be given tasks? What's the format of their uh product backlog items, their work items, their tickets? Emails, whatever. So it'll be perfectly formatted in the way that the person wants to see it in the system that they want, CCing all the or at mentioning all the people who care about it, the product owner, the Scrum Master, whatever, the tech lead. And then it has the link to the so as well as explaining the issue, giving reproduction steps, all that sort of stuff in text, it also has a link to the video. So they can see it from the horse's mouth. Because we did this in 2004. And at the time, AI was pretty good, but it would still sometimes go a little off the rails. Like you would explain it weirdly and it wouldn't quite get what you're saying. And so the written stuff used to not be quite spot on sometimes. So sometimes they they take that as a guide and then they'd watch the video for what you really said. But now, honestly, I barely watch the videos now because the AI is so good at cutting through the guff and just writing down exactly what you want and why, it's amazing. So this this is, I was gonna talk about this later, but this is this is a thing in general. If you're building a product with AI, if you are hoping that the AI doesn't get any smarter or your product becomes irrelevant, you're building the wrong product. It's gonna get smarter. If you're hoping, if if your product is almost good, and if only the AI was a bit smarter, it would be great. Then you'll be first to market because you've already built it, and by the time the AI is good enough, you've got the perfect product.
SPEAKER_02Yeah.
SPEAKER_01So that's I I think we're on the right side of that, and that's really exciting.
SPEAKER_02Yeah, I love that actually, because you're sort of not reinventing the wheel. There was a way to do it before. Like I remember having to record myself on Videard and then download that and then put it into the script, get the transcript and all sorts of different tools, and it took forever. You've just found a really easy way to do everything and with the AI, like add an extra step on top of that, but it just makes it so much easier and so much more productive.
SPEAKER_01So well, one one part of that is um the the best UI is no UI, right? So it was always like, okay, what's the best way for us to figure out like what what user interface do people want to be able to put this in? What's the easier way to do it? Ideally, it's it's something called ambient intelligence, where it's just listening. It's like a Google Home at home, whatever, right? So it's just it's just there in the background listening. And when it's needed, it just interjects. Or even in a perfect world without even being asked, it just knows you well enough, right?
SPEAKER_02That's that's giving you random ads laid.
SPEAKER_01Exactly, yeah, exactly. Yeah. So that's a that's a future thing. But um the whole the whole like the best UI is no UI. I really like that. I really like the idea that I almost have like a little helper sitting off to the side and they're just doing the things that I'm tossing their way and they're not bothering me with anything. And I can just speak the same way I I would speak if I wanted to give someone a task at work, I would just treat it as another employee.
SPEAKER_02And it's a it's a great product because it solves a lot of problems. But how did you get it out into the market? How did you market it? I know you've got a TV platform as well. So was that one part of what?
SPEAKER_01Uh look, it was tricky. Um when we built it, we built it for ourselves first. And we built it with all the systems that we had in place. We're like, all right, well, we have dynamic CRM um for our products and our people. So that's great. We'll we'll work out like if I say Bob, who's Bob? Oh, Bob Northman that works in this department. I know their email address, I'll send them an email. Um we had all our projects. Like you you had to be able to say, hey, I found a bug in Yakshaver, and it had to know what Yakshaver means, right? Or I've there's a problem in the website. What's the website? Who's that team? So you needed some mapping tables to work out, to give context and jargon and all the sort of things that you might say and you wanted to understand. We brought that from Dynamics. That was a mistake. Not everyone's using Dynamics, right? So later on we built that separately into the product directly. Um, we for for it to feed back to us, like asking any any um dis disambiguation questions, like, hey, I think you said Yakshaver, but did you say something else that it also sounded a bit like which project do you want this in? Or hey, I've done it. Here's here's the the we call it PBIs, product backup guidance. Here's the PBI. Oh no, you've done it in the wrong place. Move it here, or CC that person or whatever. If you wanted to make changes to it, it needed a way to communicate with us. So we used Microsoft Teams, because again, we're a Microsoft house, that's what we had. Uh, we used GitHub and Azure DevOps and Zendesk for our task tracking. All these things became dependencies. So when we tried to go to market, we're like, hey, are you using Teams, Dynamics, GitHub, you know, Zendesk and Azure DevOps? Then we've got the product for you. But not everyone's using that combination, right? So we had to strip out dependencies uh when we went to market. And that was a big job. Um, and it was a job we had to do before AI was good enough at coding that it could do a lot of that work for us. Yeah. Um, so yeah, I think I think if we were doing it again, we would think of how would somebody who's not us do this rather than how should we do this first? Interesting.
SPEAKER_00Yeah.
SPEAKER_01Um, like maybe do a bit of market research or something. We always think of ourselves as the primary customer. We think if something's good enough that we love it, then everyone else will love it. But also we live in a specific ecosystem, right? So stripping out teams away from it, that was tricky, that really tricky. Uh also the permissions required inside Teams are very um not permissive. Like it was very hard to get it to the point that uh it would work well. And for for ourselves, we'd bend over backwards and just say, yeah, let's do whatever it wants. We're not worried about security. We own the thing, we know we're not being malicious. Other companies don't know we're not really being malicious. So we had that they would only open up the permission doors so far, you know, of what they're comfortable with based on how much they trust you. So that was that those were the two biggest problems we had with our go-to-market.
Go To Market Lessons And Dependencies
unknownYeah.
SPEAKER_01And required a lot of redevelopment.
SPEAKER_02Well, no, it it sounds like just putting my marketers' hat on, it sounds like you sort of did segment them in a good way where you can do some hyper-personalized messages. So you said, all right, these guys are using Teams, therefore, this is the message we're gonna go to them with. These guys are using dynamics, therefore, this is the problem they have and this is the problem we can solve.
SPEAKER_01Um well, it the problem was these were just um places that the data lived. It wasn't problems they had. If they live in that ecosystem, if they're already using Teams, then sure, it's nice that it'll talk to you in Teams. Like that's where you're talking to people anyway. Um but if you're not using Teams, then you're blocked altogether, right? But yeah, no, that that was nice. Um, I think one go-to-market strategy we took, which was uh, I don't know whether it was a good one or not, but it's we did it. Um when we first built this, we got very excited. We're like, this is great. You can give anyone any task. This isn't a developer tool. This is, hey, I want to give someone a job. What's the best way for me to give them that job? Boom, we can use this thing. Like for building inspectors walking around with their iPad looking at you know things that are broken on the wall. Look, there's a crack there, you know, whatever, and it going into the task tracking system. Um but it it's, as you know, hard to sell something to everyone, right? So we made a niche. Yeah, you need a niche. So we made the decision to call it a developer tool.
unknownYeah.
SPEAKER_01You know? Um and it it reduced our town, our total addressable market substantially.
SPEAKER_02For
Marketing With Automation And Brand Risk
SPEAKER_02all the CMOs listening and they're wondering how can we use AI to move the needle on the pipeline, what can you say to them?
SPEAKER_01Okay. It depends what industry you're in, right? We work in AI, we can get away with some AI smell in our marketing materials, right? So for us, the sky's the limit. Anything AI can do, we should be doing. And if we're not, then, you know, what are we doing? Right. So so I um I run a marketing team as well. Um, and my my thing to them is always, you know, anything they tell me they did that week, I'm like, right, did you save that as a skill? Do you have to do that next time, or is that going to be taken care of for you? Uh so work a little slower, learn the tools, automate everything, um, and you can punch way above your weight in marketing. However, if you are, if you have a lot of brand risk, um, if you are, you know, you you will be judged, for instance, if you're working in entertainment, right? There's there's a lot of um negative sentiment about AI use in filmmaking and voiceovers and all that kind of stuff because people are losing their jobs and all sorts of reasons. Um, so if you are using AI, you should be using it in the back office to do your research, to help, you know, automate uh repeating things, but you shouldn't be using it for content at all. You need to be super careful that your all of your copy is handcrafted, lovingly made, a slight like a rogue M-dash could ruin you, right? So it does depend on your on your industry.
SPEAKER_02Yeah. No, I love that you you said that. I think um I've given myself the job title head of automation because I agree. I mean anything that can be done that is repetitive, you got to automate it and focus on other things. But I think people are scared of automation because every time I bring it up, I'm like, you can automate this. They sort of go, uh and I wonder if that's because they're scared of losing their job, they're scared of things changing, they're scared of, well, if someone else is doing this, like a machine, then what am I doing?
SPEAKER_01Yeah.
SPEAKER_02Um are you finding that come up a lot?
SPEAKER_01Yeah. Uh and
Agents Can Do Almost Anything
SPEAKER_01there's there's layers of that, right? So there's um, for instance, there are computer using agents now, CUAs, and browser using agents that can do things uh that we I work in software. My job's always been automation, right? So so we always are building systems to replace paper-based systems or or human systems with with something more efficient, right? Um up until just recently, we needed an API to put data into any system, right? We couldn't, we couldn't do it without that. If you didn't have programmatic access, sorry, you're gonna need a human to do that for you. Now you've got MCPs that make that a bit simpler, um, which means you have more computer access to things. But then even if you don't have that, you've got CUAs and browser using agents, which can literally take care of your computer, open up a browser, click, click, click, type, type, type, done. So literally, and I mean this literally, anything you, any work you do on a computer or I do on a computer, I could get AI to do on a computer now. Right? No exceptions, right? The only uh risk is where there needs to be some communication happening beforehand, or you it's non-deterministic systems. It's like uh getting an employee and saying, hey, can you do this important work for me and hoping they have all the context to do the job right? That's the only risk. The the risk is no longer, oh, it can't do it. It can do it, but what will it do? Will it do some unintended things on the way to doing what you want it to do?
SPEAKER_02Yeah.
Human Judgement Context And Work Slop
SPEAKER_02I find that AI can do most things, but where do you see we should still preserve that human element? Like obviously, you know, when we're talking to clients, it's then they're gonna want to talk to a human. They're not gonna want to talk to an AI. Um where else do you see the human element is?
SPEAKER_01So um at the moment, it's still not strategic. It's tactical, right? It's taking care of the day-to-day. Um, if you're using it uh more zoomed out than that in a strategic level, it's you're probably early. It's probably not quite ready for that. It it can make a shot at it, but it doesn't have our shared life experience. So we, if I'm talking to you, I say I'm saying you work for me, I'm talking to you, and I'm gonna give you a job. I know that there's a lot we agree on before I open my mouth. There's a lot of shared heuristics and shortcuts that we take, right? We both probably don't kick dogs, right? Or whatever, right? There's there's a lot there's a lot of things with our shared humanity that gives us a foundation to build upon, right? Plus, we're both living in Australia, we both speak English. Like, there's a lot of assumptions I can make that you will understand already. We don't have a shared living experience with AI. Um, so sometimes what you would assume that if you gave to a coworker that they would just know, you give it to an AI and they do something silly and you're like, well, why do you do that? Right? Because you assume that they would know everything you know. So that's one thing. Be careful. Context engineering is a thing, right? Make sure that the AI has all the information that it needs. Be weirdly explicit in what you uh want something to do. Um have you heard of the concept of work slop?
SPEAKER_02I've heard of AI slop.
SPEAKER_01So AI slop is, you know, produced with AI, chucked on social media, looks rubbish, right? Work slop is the work variant of that, right? So I need to send an email to a coworker, but I just put it in Claude and just output whatever whatever Claude says. So stuff that's overly verbose or roundabout or just just not to the point, right? Um yeah, it's known as work slop. It's basically using AI slop in your work. Uh that's you've got to be really careful of. So I I made a video about this a little while ago about being a um, what is it, half horse, half human?
SPEAKER_02Centaur.
SPEAKER_01Centaur, being a centaur.
SPEAKER_02I'm a Sagittarius. Oh, there you go. Okay, very good.
SPEAKER_01So being a centaur is like being half human, half AI, using the tools correctly. It's it's don't just say, hey, AI, what would you do and then give the answer. First off, give it a few bullet points of what you think the answer should be, then give it to AI, then take that output, use your human discernment to work out, right, that's a bit off topic, that's gonna work, that's not, uh, this is way too long, whatever. Uh adjusting it to the use case that you need and then passing it on. So your input, your output, and AI in the middle is a thinking partner.
unknownYeah.
SPEAKER_01Yeah. For for most actual content that you need to put out there. If something's more simple than that, uh sometimes AI is not the answer. So AI is non-deterministic. Uh, if you give it the same job five times, you get five outputs. Yeah. Right? Um, so sometimes you actually need a good old-fashioned, rules-based, you know, simple software system uh in conjunction with AI. So a lot of AI skills that we're building for Claude now have a scripts folder. So they do part of the job deterministically and part of the job creatively, and then they put that together.
SPEAKER_02Yeah. No, I agree with that. I think it's good to yeah, sort of don't outsource strategy to AI. Um, try and preserve some of that. And yeah, treat it in a way that it it educates you as well, not just, you know, you you get it to do the work slot and then you don't even know what you've delivered. You sort of just pass it on and forget the next day.
SPEAKER_01Well, I think owning the output is a big part of it. I think uh from a um a leadership point of view, you've got to make sure that you want your staff to use AI, but you want them to own what the AI produces. So the answer of like, hey, what does this mean? I don't know, AI wrote it. That's not an answer that's acceptable. Right. It should be, oh, I missed that in my testing. Sorry, you know. Or yeah. Basically anything that I'm producing, if I'm sending it to someone, I need to own every word that it says.
Hiring For Curiosity And Best Practice
SPEAKER_02As a COO, do you do a lot of hiring?
SPEAKER_01Uh yes, I do.
SPEAKER_02So what uh take me through some of the systems in hiring? Because I know you've made a lot of mistakes, just like all of us have in with the early days of hiring. What are maybe some crucial questions that really help you get to the, oh, this is the right person, or maybe not?
SPEAKER_01Yeah. Okay. So I think working at SSW, you're going to succeed if you're really passionate and curious about what technology could be capable of. Right. So uh a simple thing is do you for for someone coming in, do you pay for an AI subscription at the moment, or are you using free AI? If they're not paying for it, they're not really using it, right? They're not using it well. Um I I I probably go a little far. I I uh am a proper, you know, AI head. Like I think this is going places fast. Um I I actually believe we're living through the singularity now, um, which might sound mad to some people, but like when you talk to people about AI and using AI in their work, you can tell very quickly whether somebody was told to use it and they had to, or they are seeking out ways to use it and seeing how exciting it is and what's changing. And if that energizes you or scares you, it's a very different kind of hire, you know. Um and that that takes us a bit to the security side of things because you can get people who are too gung-ho with this stuff and they just like, you know, do anything it can do and and to hell with the consequences. So you've got to be well-rounded. Uh with hiring, we we we have a very slow process. Um, we have like a screening call, then we do a communication test, then we do a technical test for every single role. Doesn't matter if it's a technical role or not. If it's a sales role, we make it less technical, but it's still a bit technical.
SPEAKER_00Yeah.
SPEAKER_01Right. Because we're we're a developer company and we want people who are kind of tech eds, even in marketing or sales or whatever. So that works well because by the time you hire someone, you've basically already worked with them for a number of hours, right? You've given them feedback, you've seen how they've taken it. Uh, you've given them specific instructions and seen if they're details focused. You've seen how they use AI in their work process because that's part of the test. You know? Um so that that helps a lot. Uh but yeah, I think being they say you should be slow to hire and quick to fire. Um we are slow to hire and slow to fire. So luckily we're slow to hire because yeah, well, I think we we generally get pretty good people on the bus.
SPEAKER_02Well, sometimes you have to give people the chance and the benefit of the doubt. Like just because they've made some initial mistakes in the early days doesn't mean they can't improve. So if the other elements are there, like, you know, they fit the culture, they're willing to learn, they're passionate, you know, maybe some of those other behaviors that don't match can be changed or thought.
SPEAKER_01Yeah. Well, we we have a tricky combination of um personality types that will do well at SSW. You need to be best practice focused, right? You want to do things the best way and you you care. It's not, I don't just do it this way because I used to, I do it this way because evidence says this is the best way to do it, which means you need to be able to change what you're doing a lot. Um, but also you need to be deeply opinionated about what is the right way to do things. And those two things are often the often a lugh hits because if you're very opinion about it, opinionated about something, it's because you've done it that way for a while. Right. So uh finding someone who can be very opinionated but also change their mind really quickly uh is a rare combination, is what we're often looking for.
SPEAKER_02I've heard as well a good question to ask, especially for developers, you know, people that you want to be constantly embracing new technologies to test a passion is um what are some of the side projects you're working on outside of work? And if they start describing all these things they're building with AI, then yeah, thick.
SPEAKER_01Yeah. Well, now nowadays with uh open source GitHub projects, most developers have public stuff you can see before you even talk to them. So before you even call them up, you can see in their resume what their GitHub profile is and see, see what they're doing. So uh yeah, I mean, obviously there is a there is a fear out there that uh anyone looking for a job is actually not really talking to hiring people, they're talking to AIs. Um but yeah, it's true. So the first round of screening is with AI or should be with AI, um, because you know you can take a hundred resumes down to the 15 you should really pay attention to. Um but just heads up, if you're a developer, get a public GitHub profile and start doing cool sounding stuff publicly because the AI see it.
SPEAKER_02Yeah, for sure.
Founder Advice On Niches And Differentiation
SPEAKER_02Um so us as founders listening, what would be the one bit of advice you'd leave them with, like be it hiring, be it product, be it marketing, scaling, what do you think is the most important thing to get a company from maybe a scale up to a mid-sized company in Australia?
SPEAKER_01So a lot of uh founders right now and SaaS companies right now are building something with AI. I think figuring out your pricing model uh is a big deal, um, but it's not make or break. You need to find something that people find that solves a real problem and that people find genuinely useful. And nowadays you've got to find something that's sufficiently differentiated from what the core large language model producers make that it doesn't feel like, yeah, but I could just do that in Claude, right? Or I could just do that in ChatGPT. So if someone with a bit of fiddling do your startup uh with ChatGPT, then you haven't really got a great foundation, especially if it's useful enough that the that the big companies decide to make that a core feature in their next release. So yeah, look for niches. The the um for instance, let's say OpenAI, OpenAI is not going to go after the specific, you know, CRM market for dog trainers, right? Like that little niche that that's not worth their time, right? But you can hyper fixate on that and you can make the thing that's you know obviously the best option for those people. So pick a specific niche that you know well and really serve the hell out of it. And just get a lot of AI into your development process because nowadays the bar has raised. Uh, like people expect constant updates to their to their software.
SPEAKER_00Yeah.
SPEAKER_01And if if you know a new feature comes out with AI and your product is an AI product and doesn't have that feature, after a while, people will be like, it's not as good as basic ChatGPT, why would I use it?
SPEAKER_02Yeah. It sort of takes me back to strategy again. Because I'm just thinking about that book, playing to win, not sure if you've read it, but that's exactly what they say. You know, you've got to find a market that you're gonna win in. Um, don't play everywhere, for example. So yeah, you're not gonna be competing with open AI, like that's impossible. But if you find your niche and you make the best product or service around that niche and you market it specifically to that market, then you're winning.
SPEAKER_01Yeah, absolutely. And I think also the the flip side of like move fast on the new AI features is also make sure that your core use case, your core workflow, your happy path is is baked in, is that is like spot on perfect. Don't get distracted by the bells and whistles. Lock that in and then work on the bells and whistles. Because yeah, it is sometimes fun to I'll give you an example from us. So we we built, when we first built it, we built it as a uh an AI enabled workflow, meaning there was AI in each node of the workflow, but the whole the specifics of this, then this, then this, then this was baked in by the developers. Right? We then switched to an agentic workflow uh after like a year. So it was it was working perfectly. Um, but we decided we wanted people to be able to um do whatever it the thing was capable of doing rather than just do what the workflow enabled. Um so we wanted it to be a gentic so that it made its own workflow on the fly. Right? It has access to the right connectors and anything you ask for it can do. So for instance, one thing our workflow didn't enable was having one video create two or three tasks in different systems. It was always one video, one task. Right. So we we spent a lot of time uh changing that and and making it like that. There was no ROI uh improvement at all. It didn't really move the needle at all. Uh in fact, I think people liked it more arguable, but I think some people liked it more when they they were, it was more predictable what the outcome would be because it was only it only had one workflow. So you knew this is how I use it. I use it this way every time. When it became too big, it like people started pushing at the edges of what was possible too much, and they ended up with bad outputs.
SPEAKER_00Yeah.
SPEAKER_01So yeah, lock in the happy path, work out what's the core use case for this thing, and make sure that that is really good before you go exploring too much.
SPEAKER_02Great
Work Life Balance That You Schedule
SPEAKER_02advice. My last question for you tradition on this podcast is if you could go back in time and give yourself one bit of advice, be it work, be it live, how far back will you go and what would that advice be?
SPEAKER_01So the the rote answer is I wouldn't go back in time because all the bad things that happened to me made me who I am today. But realistically, I think I would have told myself to buy Bitcoin really early.
SPEAKER_02No answer.
SPEAKER_01Um Yeah, look, I I think I'm just thinking. It's actually it's actually a very hard question to answer correctly. Uh I think there's a right answer. There's no there's no right answer. I look, I think work-life balance is really important. Um and often people get uh tunnel vision on what they're trying to achieve, especially when something exciting like a new a new startup, like the actually everything happens and they they go all in, uh, and then other parts of your life can suffer. Um I think, you know, you you want anything you do, you want to do well. There's no point in doing anything, not doing it well. But but find the balance that you want in your life and almost write that down and then be true to that. And you'll find that the like your your whole life will be better as a result.
SPEAKER_02I love that. They say um you can have everything, but not all at once. So it's just about picking what am I gonna focus on in this phase of my life. So you can't have, I think the the time where you feel stuck is when you want two competing things at the same time. Like I want to be at the top of my career and I also want to be the greatest mom or dad. Yeah, that's not gonna happen. So you've got to kind of pick one for, you know, maybe until they're like 20s and ignoring you, and then I'll focus on my career. Um, but yeah, you kind of you do need to stay focused on one thing, I think.
SPEAKER_01But actually, no, I I almost take the opposite view. So I think uh you can if you dedicate time in your day or in your week to any one thing, it's almost like that cognitive offloading thing that we talked about earlier. You you can be a fantastic parent in the moment, and if you dedicate enough moments, you can be a fantastic parent for what's acceptable, right? You can be fantastic at work while you're at work. It's when you get the crossover. When you're working when you're at home or you're dealing with home stuff when you're at work, that's when you're gonna have those dramas. Yeah. But no, I think I think I've always been a big believer in getting that right. Like you need you need physical, social, um, monetary, like cognitive, like you need all these different things to be a well-balanced person. And you've got to dedicate time to each of those things. They won't just happen accidentally. So you kind of have to be a bit uh anal with your schedule.
SPEAKER_02Yeah, habits.
SPEAKER_01Yeah.
Final Thanks And Wrap
SPEAKER_02Well, thank you so much for being on the show. No worries, it was great. Thank you very much.
Podcasts we love
Check out these other fine podcasts recommended by us, not an algorithm.
