Ask Us Anything: Negative experiences in business deals + Would you rather Human or AI phone agents?

0:00I'm Adam Schwab. I'm Adir Shifman. And this is The Contrarians with Adam and Adir.

0:08And we are back. Episode 182, Ask Us Anything. We love getting your questions. We have the best listeners in Australia, nay the world. So we love your very intelligent questions. What are you channeling a horse? Over to you, Mark. Question one this week from Terence Tam. Hey, Adam and Adir, if you're doing a deal with a reputable company, whether they've come to you or you're pursuing them, but the person you're dealing with is unpleasant and shows questionable character traits, do you push through anyway? Does the strength of the company and the opportunity override your reservations about the individual?

0:50Where do you draw the line between business first pragmatism and doing business with someone who potentially has poor values? That's a great question. A very idea question this one. So I've had these kind of experiences. Like I'll tell you what it depends on because it depends. But the baseline for me is if I'm ever going to be involved with someone in anything, and I think business is a subset of anything, then that person's ethics is probably the first thing that I'm interested in. Like I'll do business with people who I think are ethical and also, you know, smart and fun and all of these things.

1:32But I think that's, that's the kind of baseline is being ethical. Sometimes you bump into bankers that are not ethical when you're doing a deal, I can kind of manage that because that's not really the fault of the founder in that situation. And I haven't had that very often, but I've had it at least twice when I think that's been a problem. And often I raise it with the founder gently and say, I don't think this is actually very fair. Like, I don't think we should be doing that. But I can tell you definitively, if I think a business is incredible and I'm planning on acquiring it and retaining the founders to be a part of the broader business, and I think that that founder has got big ethical issues, then I'm not going to be completing that deal.

2:12And I'd go further and say, there's only been two occasions in my entire life where I've been involved with people where at some point in the relationship, I've thought, oh gosh, like that does just not feel like ethical behavior. I've got question marks. And both times, because things were going so well, I ignored it. and both times they ended in catastrophe for me in one way or another. And I would never ignore that again. I would get the hell out of there if that happened again. Yeah, I wasn't quite sure if Terrence Mann, as a sort of investor founder relationship or as a kind of counterparty commercially.

2:48That's kind of the impression I got, that you were trying to sell to somebody. He was almost talking about if you're selling to somebody. So I think there's so many different… Well, who cares about that? Yeah, I think there's so many different categories. I think if it was an investee, So if you're investing behind a founder and they were of dubious moral quality, I think that's just an absolute no-go. Like, well, investors have so many more options than investees. I think you just move on to someone who's honest. I think that's a pretty key requirement. In so many ways, a founder can screw an investor over and vice versa, to be honest.

3:19I think you've had experience with both sides of the coin, investors. I've had experience with investors, not many, thankfully, just one or two, who are of extremely dubious character. and it certainly rebounded on us and I would never do that again. That was a huge lesson. You're much better off not investing in someone of dubious character than taking the money or whatever the situation is. You're absolutely right. Yeah, I totally agree. I didn't think about that. Yeah, I totally agree. Well, there's a hierarchy of difficulty with unethical people or people that you think, I don't want to say unethical, people whose behavior makes you think that's just so obviously the wrong way to behave that like that makes me very uncomfortable about the individual and the hierarchy is the easiest is if you're investing in someone you don't invest in them so that's easy right and then it gets more difficult like well the easiest is if you're selling to someone who's unethical as long as you get the money and you're selling out entirely I actually don't think that matters.

4:21Like that's their problem, right? And then the next level up is investing in someone, so you don't do it. And then the next level up is probably buying a business and you want to keep the founder and it turns out that it's questionable. I wouldn't proceed with that transaction. And then the next hardest is maybe, I can't work out if you're in business with someone already and it's going well and you find out they're doing things that you think are unethical or an investor and you need money and that investor is unethical. Yeah. Would you have not, were you in a position where you had to take the money or die?

4:54We actually never raised money until well later on the piece. And when we did raise money, we raised it from a very ethical fund and very ethical person in P12. So we've never had at Luxury Escapes that issue. We've had other shareholders in this business, other businesses that I'm sort of somewhat talking about without sort of giving it away who I'm talking about. But funnily enough, when I've invested in ethical founders and I've invested, and a A few founders come to mind and a bunch of other great founders, but Matt Dyer from Bookwell and Eat Now and obviously Nathan as well, Teresa April from Brand Crush and Bart Jesman from FinSuite, three slash four outstanding ethical founders all delivered us a great result and there are other people who maybe weren't as ethical, could have been great founders and smart in other ways, but delivered a terrible result.

5:41And there are some problems you invest behind ethical people who are also a bad result, so it's not a guarantee, but certainly I think you increase your chances of good result if you actually invest behind someone who's a good person. I think when I look back on taking investors' money, people that thought these individuals were subsequently very unethical, if I could wind back the clock and not take their money, almost irrespective of the consequences of doing that, like I would have figured it out somehow, I wish I could wind back the clock and not take their money and never have had anything to do with them as people.

6:13Yeah, I couldn't agree. I think, and we haven't had this specific situation, but I 100 % would not, if the situation came up, don't take the money. I 100 % agree. Just bootstrap. The problems you have with unethical investors who are willing to rip your face off at every opportunity. Because remember, when you sign a shareholders agreement with somebody, they can pretty much veto anything and blackmail you at any time because you almost always need, yeah, you can drag people in certain circumstances, but you almost always need someone to sign off on stuff down the track. That can be one time, can be many times.

6:45So assume that you'll need someone sign off on stuff and assume that person can blackmail you or greenmail you if they want to. And unethical people will is the problem. So I'd be very careful about that. That's a really good question. We can move on to question number two. Question two this week comes in from Corneliu Pop on LinkedIn. Corneliu asks, quick question for the Ask Us Anything episode. Does Adam's team at Luxury Escape use a SaaS platform for managing seller slash partner accounts? Or is it a custom built solution? If custom, any idea which language or stack it's built on? Also curious how they handle commission calculations with their sellers.

7:28Is that baked into that system? Or is it a separate tool? And Adir, obviously it was directed at Adam. But if there's anything you can offer, we'd love to hear it. What a great name, Cornelio. Um, great question. So we use Salesforce as a CRM for our, our clients. We used to use, so Salesforce has a number of what they call clouds. So their original legacy business is called sales cloud, which is basically managing your customers in our case, hotels and tour operators, but it could be, it could be any kind of customer really. And then they've got also marketing cloud, uh, which effectively helps you send your emails and your SMSs, that kind of stuff.

8:03We've, we got rid of that move to Braze, which is, which is a far, far better product. We've also got what's called Service Cloud and we've moved to Zendesk. And Service Cloud manages effectively your tickets, interactions with customers. So we have kept Sales Cloud. It is really the leading product in that kind of field. There are some competitors. I think Zoho is a few competitors of Sales Cloud. I think within five years, we'll have coded our own version of Sales Cloud. I think it's a big undertaking. It's not simple to move off it. But that said, I think the ability for us to customize one for our needs.

8:37The problem with Salesforce is you need to have multiple, and we've got multiple exceptionally good Salesforce developers, but you need Salesforce developers that work on top of the Salesforce platform. So it's not only expensive, you pay a per seat license fee to Salesforce, you've been really going to spend another 500, 300, whatever grand is on people to manage the platform. So Salesforce is expensive. In terms of what we code it in when we do code it, that's probably a question that's above my station. I've asked our head of engineering and our CTO about that. We use sort of React Native and a bunch of stuff.

9:10But so I'd have to defer to them on exactly what we'd coded in. But you'd presumably code it in your own. You'd add it to your admin platform. So we've got an admin platform that we manage all our sort of partner information details. In terms of how we calculate commissions, that's actually using multiple things. So we essentially get, we've got a Tableau sort of data lake essentially. So we get the data out of Tableau and then we use Excel to calculate the commissions. Our commissions are relatively complex. Like most commission calculations, there's hurdles and gates and all that sort of stuff and multiple different ways that the people earn commission.

9:46It's not just a revenue-based commission. There's a bunch of activity stuff in there as well. And if you're on the phones, there's a bunch of conversion metrics. So it's sort of another conversation. But no, we don't use Salesforce to calculate commissions. I think you'd struggle to use Salesforce for that purpose. I think a lot of companies still use Excel and then you've just got to get the data from somewhere. So we've got a pretty big admin backend. That's what most of our developers work on. We've got 150 developers, engineers, design product who are working on this backend. It does link into things like Salesforce, but ultimately Salesforce is a plugin.

10:20It's got a specific purpose, Sales Cloud. It's fine for now, but as our dear nose, I can't wait for the day where we turn this off and I think it will get to the point in 10 years, I think most people will have coded their own, but I think that time is not imminent for Salesforce. I think they've got a pretty good moat. I know nothing about this world. Like I know these names and I, like this is all operational. This is not my world, right? I don't know what's going on with this stuff. And so people that are much smarter than me at this and much better, like they handle all of this. You know how dubious I am on this claim of like you'll code it all up, but like can you break out the cost of this part of the platform or not really?

10:57Like what do you think this costs you? Like how much will you save if you go to yourself? Not including all of the extra costs and those sleepless nights for all the bugs that you introduce into your own system, but like what do you think you'll save dollar wise? Well, we pay probably 500 grand for Salesforce, just sales cloud. That's before we pay the people to manage it as well. So expensive. And that's a perceived basis. So you pay for content writers who might use it for a part of it. You pay for obviously your salespeople. So you pay for a bunch of different people. And also the beauty of coding your own one is it's your own one.

11:27So we can make your own changes to it. You can customize it for your own purpose. Salesforce, you're kind of retrofitting into your platform, which is a bit frustrating. Whereas if you have your own one, you can effectively build it yourself and it's part of your admin. So it's certainly preferable to have your own. Historically, it hasn't been worth the investment. You're better off using the off-the-shelf Salesforce type product. But clearly, we're getting the cost of software. And this is, I guess, the AI point is AI makes software a lot cheaper. You can create a lot more with less. So it comes to the point where that balance between using third-party software and creating your own becomes a much more different calculation.

12:04So when you look at why the SaaS catastrophe or SaaS collapse or SaaS crash, whatever you want to call it, a lot of that is not necessarily happening today or tomorrow. It's coming in the terminal value of these businesses. And in two, three, five, 10 years time, a lot less businesses will need these plugins that have been coded over 10, 15, 20 years. Like Salesforce is an extensive product, but it's a terrible product people hate and it's not fit for purpose. You've got to pay people to do it. You've always got to use consultants to get it set up. It's a really bad product. And I think businesses can't wait to get off it.

12:33They just haven't been able to. So if you think about the economics of what you're proposing, you're not going to go and use co-work to write this software. I mean, definitely you're not personally, but no one is. But maybe they can use Claude Code if they're a good engine. Claude Code, yeah. You use cursor and Claude Code or whatever else. And so let's say it's a good engineer. You need a good engineer to do this because, you know, I've said this before, coding is not software engineering, so you need a proper software engineer, and they'll use that code. And let's say you can get away with one.

13:00Now, this one, this is the person that you're going to be calling at three in the morning when there's a bug or something's down. Wake up, there's a bug, no one can use the thing. And I know, you know, like the big company SaaS businesses, like they provide 24 seven support, probably. I would, I hate to break it to you, Salesforce, I'd answer the phone call, I'd answer your phone call at 2am. So you might take two days to get hold of someone there. Well, for 500K, they should be providing with good support, frankly. It's shocking that you pay 500K. Like, I'm saying this in a very respectful way, but you're not a big business.

13:29Like, you're a big business for a private business that's founder-owned and run, but like, you're not Telstra. We used to pay like 1.2 million when we had multiple clouds. This is, or we reduced this because we got rid of service cloud and marketing cloud. We pay other vendors not that different amount, but yeah, we initially used to pay Salesforce like over a million dollars a year. So you have this person, all right, they're not answering the phone, whatever, that's fine. And then, I don't think it's fine, by the way. And then, so you're going to have to hire someone because you can't just dump this on your existing.

14:01But we already have multiple people working on Salesforce. So you're, hang on, developers, like high quality engineers. Yes, Salesforce developers, they're called. Okay, so you're going to go and repurpose them. Well, whether you repurpose them or they do other stuff, who knows, but you have an existing. So do you think you're going to have no net increase in your employee expenses to build your own software? My gut feel is, and this is just a guess, I don't know, but I suspect we'll save annually, we'll save a significant amount of money, like as in the initial build. And then once you've built it, the upkeep's going to be less.

14:35And it's going to be more customized to us, so it's going to be better. I think it's going to be better in every way. But do you know what you're going to do if it's really good? You're going to say, people are going to ask you this question, and then you're going to have this conversation. Should we make this available for other people? That's what you're going to say. I don't think it could because it would be integrated with our app. It would be part of our admin back-end system essentially. So it's not the kind of thing you'd necessarily be able to sell easily or not the kind of thing we'd be able to easily white label, nor would we because that's not the business we're in and these businesses aren't being valued as highly as they used to be.

15:05So I think the realisation will come that businesses can start coding a lot of this stuff themselves. I think that's why I think the question is so good. And we're not that far away. I think for some things, and I think you said it brilliantly, what was about four weeks ago on our episode, which SaaS businesses are in trouble. And if it's a complex, like that whole WiseTech stuff a couple of weeks ago when WiseTech dropped 20 % in a week, that was because someone vibe coded some big chunk of WiseTech. I think WiseTech will actually take longer to vibe, to code. And people tell me that people in that industry, I've heard, because they jacked everyone's prices recently.

15:43people really wanted to get rid of them, right? And it's hard to get off WiseTech. Like it's deeply embedded in people's systems. I think the other big issue for SaaS, and you actually said it there, is where the best part of SaaS is the switching costs. So Atlassian and Xero and WiseTech don't have particularly great – none of these products are particularly great. I'd probably say Xero is the best, then Cargo Wise, then Atlassian. And Xero is a good product. Yeah, and Xero is a good product. But they're all very sticky. And Xero actually is cheap and a good product. It's just badly run now as a business, but it's a great product.

16:15But they're all sticky and they keep putting that price up. And as you said, Cargo Wise keeps putting its price up. Wise, it keeps putting the price up. Eventually, you piss your customers off so much that they go, stuff this, I'm going to spend whatever it takes to get off these criminals. So you can only push. That's another reason why these SaaS businesses are being written down because they're losing pricing power. So everything about these businesses, the terminal value business, is actually in a lot of trouble. So everybody says, oh, don't worry about SaaS. It's going to be fine. I actually think they're wrong.

16:39I think these aspects are in all sorts of trouble. It's interesting that when someone increases their prices and you call them a criminal, that's a low bar. But what about if I raise this last point? So if Salesforce dramatically improved or increased the value of what they were providing to you for the same price, then you'd keep them because then you'd say, well, we could go and maybe it would cost us a bit more, maybe we'd save if we did it ourselves, et cetera. But actually, they're now doing all of this other stuff and adding all of this other value. Or it got cheaper. I mean, that's the nightmare for Salesforce, right?

17:16Is prices dropping. I think the problem Salesforce has with you, and I just want to, I think it's a very important distinction. I would veer it a little bit with this. But like, I always say you're the worst customer because you're too smart. I never know if that's a compliment or an insult. No, it's a compliment. It's a compliment because you're too smart and you're too across the detail. and like you do not suffer being shafted for very long. Like it really, you hate the injustice of it and you're prepared to suffer some pain to get out of that injustice. No, it's definitely a huge compliment, but like that is a tough customer.

17:49And so, and the other problem that you represent for a Salesforce is effectively you're a tech business. And so you've got all these developers doing all this stuff as it is. And most businesses that Salesforce is selling to, they're not tech businesses. they're just using salesforce as a platform to run their business and so i think that maybe i don't want to concede this point yet but it's possible that you'll definitely have a dig at coding your own thing and maybe you'll even succeed but i think using you as an archetype of the risks that this industry is facing might not be 100 right because most do not have your sophistication and they're not tech businesses i think i think the tech business point's a really fair point it's hard for if you if you just don't have tech resource or you got one tech one product guy in india it's gonna be pretty hard to vibe code a sales force well i think what you might experience though is sales force having to drop that per seat costs potentially uh for everyone well they don't even have to drop their per seat cost the big problem they're going to have yeah is like if we assume that either um things are going to get a hell of a lot more efficient with the same number of technologists or there's going to be a reduction in developers i'm not sure which one it is but either way like they're gonna they're certainly not gonna be able to put up their costs per seat so i have a feeling what we're mostly gonna see and this might influence your view as well well you just care about the net cost at the end of they don't care how they bill you what they're gonna figure out is how to make the same amount of money from telstra and you'll just be downstream of whatever that decision is you know what i mean like they're not gonna worry about you you're just downstream yeah it'll be interesting to see and so just because i was veered from the question but like are you saying to someone now you should go and build your own today or not because that were really kind of the question was alluding to that probably not quite yet i think would be there i think it was really if you look at braise which is effectively a marketing platform and how we send our emails and our sms's we actually have zero intent to code braise ourself because it's much more technical and b they're doing a great job and C, they're not as expensive.

19:54So I think to your earlier question, it was a good question. I guess because there's so much animus towards a Salesforce specifically that we'd be keener to move off Salesforce. And I think the upside of moving off, I don't think I can do a better job than Braze does for doing what Braze does, whereas I'm pretty sure we can do a better job than Salesforce does for what Salesforce does. So I think it really depends on the utility these platforms are giving the customer, which is really impactful on the want of the customer to build themselves. Very interesting. We'll go to a super quick break back with question number three just in a moment.

20:32And we're back, Michael, question number three. Final question for this week from Duncan. He asks, would you rather call a business and speak to a Gen AI agent or a human if the wait time for the human was between four and six minutes? your answer is important to us oh that's a funny end of that question i like that a lot that's a that's very good i just want to speak to no one that's my first desire i don't talk to anyone and things are going to go wrong but just let me fix them using whatever platform you've made available but if you can't do that if you have to speak to someone um then i just want to speak to someone exceptionally competent in a reasonable amount of time, four to six minutes.

21:22I mean, a lot of industries, four to six minutes is about one-tenth of the amount of time that it actually takes to speak to someone. So I don't care about four to six minutes. We can make this question. Would you rather go around in circles with a D-grade Gen AI agent or wait 40 minutes for a human being who is offshore and from Australia? And that's a trickier question. I think in that situation I'm all right with trying the agent. Like the problem with a lot of these agents, you must have had people send you stuff. I don't know, for some reason people think that I'm really interested in their problems with AI agents and they send me like the little screenshots of their like absurd interactions and it's like I don't care.

22:08Like I don't even care when you tell me. Some sort of weird dick pic thing? Version of weird dick pic thing? You're not the Luke Sayers of IT. No, I'm not the sender. I'm the receiver. Yeah, true. And so. Poor girl. Like you know when you talk about your Uber problems? Like I struggle to stay awake for those. And like I definitely can't look at other people's problems with voyeuristically your AI issues. So they are pretty hopeless, a lot of them now. I know if Laura Keats listening, they're going to post something and tell me how hopeless I am for saying that. But basically I think a competent AI agent can handle, I'm going to make up a number, but like 80 % of what people are generally asking for because people, like there's a huge.

22:54So you're talking about a voice agent or a text agent? Oh, the voice agents are awful. Yeah, they're awful. Yeah, I'd much rather just text to an agent than hear this like ridiculous voice, right? In fact, I hate those voices in video games. They put them in video games now, these AI voices. It's so obvious they're not real. And so I'm all in favor of AI agents. I just think the CEO of whatever company it is that deployed that agent, they should be forced to spend like two hours using that agent to solve their problems before it gets deployed onto their poor customers. and if the CEO says, yeah, this is actually a really good experience, I'm all in favour, I think the world will shift to AI agents by and large is my guess.

23:40So we're deep into this so I can give you our personal experience from the company side. I'd also give you my views from the customer side but from the company side, so we are using Lora Keat. As you mentioned, Lora Keat's Australian business backed by Blackbird, run by Steve. Now what that Lora Keat post on LinkedIn is going to be is, oh, I listened to the contrarians. Adam's such a fantastic guy. Adir, he's totally clueless about AI. So I think we've had a journey with Laura Keat. So they took over from another, an Israeli business who was struggling a bit with us. And Laura Keat, to that credit, they actually do a lot of work with him and hers, one of our company we've spoken a lot about on the pod in the last few weeks.

24:15But they've probably, we were asking for them to do some complex stuff they hadn't done before. And they've actually made some pretty good progress in the last couple of weeks. So we are actually trialing another company alongside them to see who can get better. But Laura Keat, to that credit, But we're down to about or up to, you can say, 90 % good, so 10 % problematic, which is obviously not as what we want to be, but we're languishing at sort of 60 % good and 40 % not good, which is not great. But at 90%, we're getting to where we want it. I think for me it was always can we get to 95%, 97 %?

24:47Then we can roll it. At the moment we're just still only using a small number of calls. It's like 50 calls, 60 calls a day. It's voice. It's voice responses. This is voice. So we've got effectively a chatbot which kind of allows you to effectively automate your questions. So do you want to refund it? So it's not really a chatbot. It's part chatbot, part kind of guided flow through what you want to do. I want to get a refund or guide you through it. But there's a huge database sitting behind this, right? Like there's all this – like, yeah. So it's kind of like before this whole everyone got excited about Gen AI, you would have said we've got this whole database and we're going to build this chatbot and it's just going to use algorithms, which, you know, is just like maths basically, right?