Why Bolting on AI is a Losing Strategy

Greg Stewart: Welcome to No Free Lunch, Africa's freshest business and tech podcast with me, host, Greg Stewart. Our topic today is why bolting on AI is a losing strategy. And with me today to unwrap this topic is Joel Rothman, the founder and CEO of of INSDI, a South African no code digital platform built specifically for insurers and financial services companies. Joel is a South African tech entrepreneur and insure specialist with over 15 years in insurance technology.

digital transformation and product strategy experience. Welcome to No Free Lunch Job. Joel Rothman: Thanks, Greg. Great to be here.

Greg Stewart: So Joel, to start off with, we hear a lot of claims with AI it's not that I'm ⁓ an AI skeptic. think are applications where it is proving ⁓ itself, a lot of claims are coming out of the AI industry. There's a gazillion consultants in the marketplace telling people this and then the next thing. And ⁓ some, claiming that costs are slashed up to 90%.

But if one takes a look at the latest reports like Simplified, Deloitte, McKinsey that kind of report that's coming out, recent reports show most insurers are still stuck in pilot phase with fewer than 15 % able to prove measurable ROI. And many admitting the sort of hidden integration and charge management costs. that have been stuck away here and there, that those costs are actually eating a big chunk of the expected gains. What's your view on that gap?

Are we overhyping returns? ⁓ is the likes of your company seeing clients achieve those kind of numbers in real South African deployments? And if so, what's the secret sauce? Joel Rothman: Yeah, so maybe, you know, a full disclosure.

I live kind of two ⁓ One is INSDI, which is advanced information gathering platform that we developed during the COVID era to help businesses get information quickly and simply from their clients ⁓ and cut the and fro that usually accompanies that ⁓ type of interaction. that's my one focus and I'm a co-founder there together with my partner Nick Eszterhausen. ⁓ been doing since 2012 was technology consulting primarily into the insurance industry. ⁓ you you mentioned, that that is exactly where, ⁓ fertile ground these kinds of AI ⁓ initiatives.

⁓ But think ⁓ the question is, there a lot of Busyness and and a lot of expensive consulting going on with very little to be shown for it and I think Up until recently I would say yes, I think to be honest. It is still a very very new field I think Whether consultants ⁓ or you? Even if you a technology professional the reality is ⁓ AI its current form has not been around for long and I think there's a lot of exploratory work going on at the moment. ⁓ know, once said to me, if you're on the bleeding edge, it hurts, it costs.

And I ⁓ we're seeing, we're seeing a lot of that. I think what is important is Greg Stewart: It's a huge pain. Joel Rothman: that the learnings do get shared almost in the public forum, the forum ideas and shared knowledge that companies can collectively learn from this and move forward fast. ⁓ There's definitely, value to gained from this.

⁓ And guess we'll get into more of what the pros and cons are of new AR world. But I think, there have been some expensive pilots ⁓ and lots of learned. we know a lot of these pilots have been shelved. There's no doubt.

Because you've got a lot of teams trying to trying to make sense of technology while adding value in. And I think that's the adding value part is sometimes lost when we get focused on the shiny things, the new technology. Greg Stewart: Yeah. And also think that there's probably a lot of pressure coming downwards from senior management that have made claims and are saying, you know, we're going to revolutionize business and at, as you say, the bleeding edge expecting things that aren't necessarily doable right now or probably doable, but it takes time to make that happen.

one to look back at things like machine learning, it's taken decades develop machine learning to the point where it is at now. And ⁓ was actually at an summit yesterday. And the point about this is that so talk about I think in some respect, there's too much talk and there is a sense of hype happening. people there were saying, well, Africa can leapfrog the rest of the world with AI.

And I'm going, but hang on a second, only about 40 % of Africa has connectivity. There's hundreds of billions of dollars shortfall in terms of infrastructure such as electricity and data centers and networks and all these other things that are required for that to be a reality. Let's not talk about leapfrogging. Let's actually just build something sensible.

And I think, you know, I like to have sort of a more grounded to this, notwithstanding the fact that there are, you know, opportunities. And I think there needs to be a distinction made between the opportunity and the hype. So let's talk a bit about the AI adaption trap. You speak about the whole bolting on of AI.

And I think this is one of those areas that is a real problem with AI implementation is people are trying to bolt things on instead of actually going back to the drawing board and going, OK, how do we reimagine this whole thing? Joel Rothman: Yeah. Yeah. Greg Stewart: do we make this AI process part and parcel ⁓ ⁓ operation rather than something we just try to plug in?

⁓ old thing, just find a plug-in and add it on, can often create a headache. Talk to us a bit about that and what you experience around that has been. Joel Rothman: Yeah, I think if there's one thing that businesses need to wrap their heads around is that AI isn't just this tool ⁓ you plug in. ⁓ a ⁓ processing tool or something that use like that.

maybe, I think the challenge here is that in your personal use, is quite tactile in a way. You can experience it quite quickly. You go to chat, GPT and bang, it comes up with something quite impressive. Now you can get PowerPoint presentations out of these AR tools.

It feels a bit magical and it's near instantaneous. I think what often happens is then business leaders think the quick solution is you plug it in and to get meeting notes automatically out of our meetings. going prepare the agenda for us. It starts taking away a lot of that busy work.

⁓ And that's think we've seen AI has been quickly adopted in ⁓ a co-pilot ⁓ ⁓ of ⁓ ⁓ Greg Stewart: Yes. Joel Rothman: where it assists with those day-to-day things. Whatever you're doing, it just takes away some of the admin for you. But I think real opportunity ⁓ for businesses to take a step back and actually say, hold on a second, if we had to start this business again ⁓ AI now being ⁓ play, what would our business actually look like?

⁓ that's it. That's a big challenge. know difficult in any organization, ⁓ small, medium, large, challenge your status quo. Number one, we know that's got to be a top-down initiative.

⁓ And ⁓ secondly, ⁓ most the especially bigger businesses that I work with, there's ⁓ a what can we call an immune system ⁓ that kicks in. ⁓ Greg Stewart: Yes. Joel Rothman: So when you start bringing these kinds of initiatives into the organization, there's a natural resistance, and rightly so. I your bigger companies, your listed companies, there's a reason that people invest in them is because they generally assure a thing.

They're not going to flip-flop overnight. is found their magic. source and they make it and that's what they focus on. I think key is when you look at the opportunity it has to be holistic and it has to really be reimagined from the ground up.

I mean I can tell you a little bit about some of the things that made me step back recently. One of them was a site called Pulseo, is AI slop. ⁓ heard of AI slop, Greg. ⁓ this generative AI can generate a lot of AI slop.

Mediocre, whether text or images, you can immediately pick ⁓ this thing was generated AI, it's very mediocre. And Greg Stewart: Yes. Joel Rothman: person who started Pulse here just took a ASLOP and reversed it. And what was amazing though is I went to the site and it just says tell me a bit about yourself and it was as simple as saying have you got a LinkedIn profile?

⁓ I said tell me about yourself it was just you know ⁓ me to something about yourself. ⁓ industry are you in? It asked just a few leading questions. ⁓ Greg Stewart: me a link.

Joel Rothman: And it went away and it said, OK, I'm going to I'm going to build a business for you. And it came back with a few options. I said, this one sounds quite interesting. It was a new insurance product.

And it said it's going to build it's going to build this. And this thing literally went away. It brought out the business plans. It brought out the marketing plans.

It it got to the point where ⁓ had a fully working. ⁓ website where I could go and buy this insurance and within let's say two hours was already saying okay I'm ready to start sending out posts on ⁓ Facebook and and that's when was like ⁓ let's put the brakes on I mean I'm just getting my head around this business opportunity ⁓ and what's Greg Stewart: Okay, well. I'm great. Joel Rothman: really interesting about it is the business model there's they've got payment mechanisms connecting with stripe and they they just take 10 ⁓ for themselves so they to get as many of these business opportunities up and running generating revenue and they're just going to take their slice from it but i think i think what astounded me though was the speed at which this thing could literally put together ⁓ Greg Stewart: Yes.

Yeah. Yeah. Joel Rothman: Something akin to a business and I know, ⁓ be fair, insurance there's ⁓ hurdles you've got to get through. There's regulatory check boxes.

You can't just suddenly start ⁓ out there and off you go. ⁓ But I've personally in my consulting capacity worked with ⁓ startups, ⁓ new initiatives and I know how long it ⁓ Greg Stewart: Yeah, yeah, Email license. Joel Rothman: generally taken about two years to get from idea to actually fully operational because you take six months to get your website going. You're trying to get your system side of things working.

You're trying to get your regulatory in order. It literally took two, on average, about two years to get that kind of business up and running. And I thought, my goodness, the one thing AI is... very capable of doing is reducing that barrier to entry right the way and think think of the key things I would want to emphasize in this discussion ⁓ is that I think if anything, there is ⁓ a ⁓ potential for a wave entrepreneurship.

I'm looking at it. Greg Stewart: Right. Joel Rothman: I know you started this conversation off with kind of the risks and the hype. And I agree, there's certainly a lot of that at the moment.

There are very real risks, no doubts, and we can go into that. That's another whole conversation. But I think it's also fair to balance it out with the opportunity. And I think it really is a call to action for whether it's...

Greg Stewart: Yes. Joel Rothman: persons who've been stuck in something of a dead end career, or people feeling like, ⁓ I can see this next technological wave is going to squeeze me out of my role. Actually, this is that opportunity to go and do something new. And I personally know three who just this year have started new initiatives ⁓ to AI.

And we can talk about that. But yeah, anyway. ⁓ Greg Stewart: us. Right.

Yeah, I think that that would be great discussion. I get going back to the original point about this whole imagining, reimagining your business. think that is such a valid point that you raised is that if you look at most a lot of businesses today, a lot of businesses was was started in an era where AI didn't exist. And so I couldn't even have been imagined as part of the structure or what it could perhaps contribute towards a business.

And that's where the risk is of the whole plug-in or bolt-on is to say, well, how do I bolt it onto my existing business? And what you said there was such a wonderful point, because I think if people aren't going back to the drawing board and saying, well, In your words, if I had to start it again, how would I imagine it? How would I see it differently? And that is, again, not an easy thing to do.

It's not a comfortable thing to do by any means. But it's probably the most sensible thing to do because it does change a lot of dynamics in terms of, you know, we're not just talking about data. We're talking about an entire organizational structure. Joel Rothman: Peace.

Greg Stewart: could be enhanced rather than just individual processes and a lot of the consultants that I speak to, that's exactly what they're doing. They're going, well, we'll take your mundane processes and just automate them for you. So you're talking about automation of processes. That machine learning has been doing that for decades.

if you had to really imagine and I Joel Rothman: Yes. ⁓ Greg Stewart: I'm also not sure that AI is yet at the point where you can say sort of intelligence ⁓ a bit of a misnomer, if I might say. because intelligence would assume ⁓ learning ⁓ cognitive learning ability and just responsive ⁓ as machine learning, responsive to certain prompts or happenings. Joel Rothman: Hmm.

Greg Stewart: I think AI does a hell of a lot better and obviously with a lot more processing power behind it. But if you had to go back and think about where this could end up in maybe a year's time and the kind of rapid development of AI and where it's heading to, one would have to sit back and provide a lot more input and sensible strategy to say, Joel Rothman: Yes. Greg Stewart: Let's reimagine this a bit. Let's let's pull this up.

And I love that idea. I love the concept. And I wish there were more people in the industry that were talking like that. let's just go on to a bit about INSDE ⁓ and ⁓ about some of successes you've had and and are people seeing a return on investment?

And and has it been because they reimagine the business? Joel Rothman: Yes. Greg Stewart: to talk to us a bit about that Joel Rothman: ⁓ so think, and I'll talk about the ⁓ before the after, the after is still a work in progress, but let me tell you a little bit about it. So, Insdi was out of...

Greg Stewart: Right. Joel Rothman: COVID, which was literally, I think, two weeks before COVID, myself and my business partner, Nick, were having coffee. And both he and I have been working in the insurance industry for many years. And we were just moaning about certain things, just the fact that...

we still getting sent PDF forms by our brokers that are not filled in, you that kind of thing. You've got to start from scratch and say, goodness, how long have we been with them? And it's got the blank form. ⁓ said, things are so clunky.

My wife had gone through a whole ⁓ journey and we were filling in ⁓ lot of ⁓ night. And most of the time you couldn't just do it on your cell phone. had to crank up the laptop and, you ⁓ ⁓ things in there. Greg Stewart: Something on the V-Scanner.

Joel Rothman: Yeah, so out of our own frustrations, and I said, well, I'm customer number one. ⁓ I want to give businesses, so INSDI standing for Instant Digital, give businesses a platform where they can really ⁓ create a guided information gathering experience for their clients. We very quickly, a client can start providing that information in a structured way, provide documents, and just simplify the journey, great client experience, and make it easy for business to then track what they're receiving through that platform.

And we had great success. built a ⁓ great client base. Our clients have stayed with us ⁓ from early days. ⁓ And it ⁓ last that we started saying, hmm, ⁓ we really want to start adding some AI features to this because if you think about it, information gathering, I think is also going to change in its nature because we, yes, we have been used to PDF forms, but let's be honest, those days are over.

I don't want to see another PDF form in my lifetime if I can help it. ⁓ But even these digital submission processes that we create, I'm thinking Greg Stewart: this. Joel Rothman: Some people would want to just talk and converse. That's one aspect.

Others would like to just upload their documents and let the AI go and fill everything in and create the structured submission. And it was with that kind of thinking we said, OK, let's build that future. was Alan Kay that said the best way to predict the future is to create it. So we said, let's do this.

⁓ Greg Stewart: to do. Joel Rothman: We started on that journey last year and something happened in December. I think it was Anthropic, you who creates Claude and Claude Code and exactly that. They released, I think it was a model called Opus 4.

5, Opus version 4.5. And our developers started testing this out. And my goodness, they could build out, let's say 60 % of what we'd taken four years to build.

60 % of that, they could build out in a two week period. And they said, look, we've managed to kind of rebuild this thing from scratch, just prompting Claude. And this is what we've ended up with. But I tell you, it was kind of a...

⁓ they say a gut punch moment, you know, where you look at this and you go, goodness, what's our moat as a business if every second company out there can go and build this thing for themselves? And I realized, okay, we're not the only business that will come to that realization when you see the power of some of the most advanced, you know, the frontier models, as they call it. And I said, then Greg Stewart: Yes. Joel Rothman: then we need to leverage that to our advantage.

We need to take our business experience. We need to take our client base. We need to then bring all of this together in a new platform that reimagines how that works so that we are one step ahead of where things used to be. it a difficult decision.

because we were both myself and my business partner were kicking again. had several quite heated discussions around this. ⁓ But we said, is the right thing. ⁓ Absolutely.

is so difficult to take a four year investment in your business and say, ⁓ we're to have to reimagine this. The world changed around us. The way the average person wants to interact with technology is Greg Stewart: and probably a couple of panic attacks. Joel Rothman: rapidly shifting and and if you don't keep pace with that you are just gonna get sidelined and that that's that's the reality that we're facing here and mean I think you saw just some my thoughts on ⁓ on that ⁓ comparing it the likes of the blockbusters and some of these ⁓ of these big companies that Nokia, for example, even, who were way ahead of the pack, saw the change and said, no, we're going to try and protect our revenue stream.

As a business owner, that's your natural instinct. I want to protect this. But I realized, if we not almost, it's a terrible word, cannibalizing our own revenue stream to explore this new world. Greg Stewart: sure.

Yeah. Me. Joel Rothman: then we're to be behind in terms of the lessons. And it's just going to take the next company to come along and say, ⁓ here's the opportunity and they're going to run with Greg Stewart: I think that's a really, really, really good lesson.

And it was just this last week where Samsung actually shut down Samsung's messaging service. You know, ⁓ you think about things like BlackBerry, I'll forget my BlackBerry, what was it? The ⁓ 866i or it was. It was fantastic.

It was at that stage, probably the best mobile phones in the world at the time. And it had that whole messaging app and all that sort of thing. also, you said, try to protect that and keep their own. And I think Apple's busy learning that lesson as well at the moment.

So need to be, I think, about these things. Joel Rothman: ⁓ shit. Greg Stewart: And I think that's a wonderful lesson. And thank you for sharing that.

But I think it's right. think your perspective on this provides a wonderful lesson for other business people in that when things shift that rapidly, you can't just play the same game. It's like the springboks going to the World Cup and playing on the rules. that we won the 95 World Cup in.

Even the players on the field and the kind of players and that sort of thing has shifted dramatically. one can't ⁓ a game with the same team. And that's great lesson. So thank you for sharing that.

Joel Rothman: Yes, I like that. Mm. Yeah. quickly, just while you're on the sporting analogies, maybe I can just throw one other analogy in there.

There's certainly an aspect of a rising tide lifts all boats. And I think certainly in terms of consumer technology and what we're getting exposed to. Absolutely, just through our smartphones, we are getting exposed to some of the best in class technology that's out there. But I think when it comes to a business, it's slightly different to that.

I say it's more like if you think of a cycling peloton. you got to make sure you're actually with that Peloton to slipstream. that's where I come back to get into the information sharing with your peers in the industry. Find out what works, what doesn't work.

That for me is the information Peloton so that you can learn those lessons together. Hopefully you share the pain, but you don't want to be a laggard at the back. Greg Stewart: Mm-hmm. Joel Rothman: you eventually gonna realize you're running out of steam and you're falling too far behind.

Greg Stewart: Yeah. And then just looking do you think that there's going to be shifts in ⁓ the insure tech environment you're in and in the insurance industry overall? And do you think that this is going to lead to real good ROI gains ⁓ to the bottom line? Joel Rothman: Mm.

think, you know, if we look back to ⁓ of the advent of the internet era, emerged ⁓ that, because distribution costs, trending down and suddenly that particular model was viable. Does it mean that a Walmart no longer exists? No, they still very established, they're very present, they did have to start thinking differently about their business. They did have to introduce ⁓ new ways distribution to ⁓ their own ⁓ while keeping their secret ⁓ and what for them.

⁓ So ⁓ think room for both. ⁓ faithless, ⁓ all doomed and if you don't jump on this bandwagon it's over. ⁓ I do think... Businesses do need to lean into this.

There are some very pragmatic steps they can take to get ready for this new wave of technology. I think you've heard it before. Get your data in order. I call it a data chassis.

Make sure your data chassis is in place because, ⁓ things are stuck in PDF documents and ⁓ documents all over the place, ⁓ think AI is going to start making sense of that. It's going to cause more confusion. AI amplifies what you've got. So number one, are you doing the right things?

The age-old ⁓ management 101, are you doing the right things firstly? Secondly, then you can say, we doing ⁓ things right? And that's where you can amplify what you're doing with AI. ⁓ the data in order.

⁓ looking at your business from a process point of view, reimagining it from ground up and actually being brutally honest with yourself. I think we can sometimes, we've got those emotional attachments to things, but being brutally honest about that ⁓ and then saying, okay, how would technology amplify this? How would it accelerate this? ⁓ And then can build on that as a long-term plan.

Greg Stewart: Yeah. Joel Rossman, think that's a marvelously sensible approach to it. And I really want to thank you for sharing some of your thoughts and insights. It's been a marvelous chat.

We probably have a few more of these, and maybe we will. But thank so much. It's been a great chat. And thank you for sharing all of those insights with us today.

Joel Rothman: Thanks, Greg. It's been a pleasure. Thanks so much.