Riding Unicorns: Venture Capital | Entrepreneurship | Technology
00:23:15 · Dan Lifshits
[00:23:15] S1: That's an amazing question. That's exactly the latter. And I think that's exactly where the difference between your typical kind of buy and build and what we're doing is coming up. Because our vision is effectively with the use of technology and AI is to streamline the entire lighting lifecycle from finding a tenant to collecting rent or property management to enhance the user experience. And you cannot do that while preserving the variety of different operational practices and having a very versatile product supporting these different ways of work. You can achieve it only by choosing one way of work and then effectively in a way pushing or suggesting everyone to adopt this way of work because you want data has proven that this way of work is better than a variety of other ways of work that exist out there. And important to note is that this is not that you defined it and that's it and it sets its tone, but it's effectively a continuous feedback loop where you test it versus reality. You buy another agency, you look at some operational process they're doing there and actually figure out that some things they invented to be doing there are smarter or more efficient or something. And then you go back and revisit your target operating model and your product supporting that for all the agencies that you already bought and fully integrated up until now. Alli is effectively the selection of the best practices of operational processes across The UK, with not just selecting the best one of the variety, but also quite often selecting none of the existing and reviewing all of them completely and rewiring your new one instead of replacing the old one. This is what effectively we built.
00:24:43 · Hector Mason, Dan Lifshits
[00:24:43] S2: Yeah. Fascinating. So, I mean, you're you're able to use data to work out and determine the optimal process across every part of the Lettings journey. And, yeah, really fascinating stuff. [00:24:55] S1: How are you... And that work, can you allow bringing most of this needed transparency and guaranteed level of service through unification, and this is critical, and through very targeted adoption and deployment of AI, integrating it into very specific predesigned workflows, where after understanding very clear process and states that this process needs to follow, this is when you then can apply AI, because effectively, it's going to be primarily just your facilitator to move the process from state A to state B, whenever you predefine know which states you want to process to follow through. [00:25:29] S2: Well, it's a great segue, because I was going to ask you how are you using AI to achieve your sort of North Star metrics? Like, what are the most exciting applications of AI and what you're doing?
00:25:40 · Dan Lifshits
[00:25:40] S1: It's across the board. I think that this is an example. The beauty of the lighting agency is that it's an example of the industry where by far most of the service delivery is communication centric workflow orchestration, where it's primarily quite labor intensive and quite repetitive admin operations that are structured around either project management type activities or communication center things, write emails, read emails, make phone calls, etcetera, etcetera, etcetera. And this is where I think it's not that it's specific activity that you can enhance, but it's rather the type of action, such as like phone call or conversation or conversational application of AI. So a few examples where we see the biggest change in adoption is, for example, on finding a tenant process to help them find the property and help landlords find the tenant is where, for example, we can increase the number of validated offers that we receive per property from one to, as usually agents are doing that, by increasing the top of the funnel, by allowing much more tenants to leave inquiries on the property and have the conversational experience with AI and chatbots and everything, and learn more about the property to the way that they can leave an application for the property, and increase the number of applications they receive per property from one tool, as traditional agencies will be doing, to like eight to 10. That allows not just to rent it faster, but also potentially have a more broad selection of the potential applicants for the landlord that benefits them because they can select a better person and a better couple. And effectively, that just makes the process more transparent for all the parties, right? Or another example might be, for example, on the property management kind of jobs resolution, where you have like a bathroom leak or something. We're cutting down the job resolution time by somewhere around 30% because kind of, yeah, I'll be surprised, but I can understand the tenants and the landlords and the British plumbers pretty well. And happy obviously to share more about how it works. It's pretty brilliant. But that allows to reduce the average job resolution time from around fifty days that is average across The UK to around thirty where we're now. And I think that we'll get this to less than ten days as product continues getting better and better.
00:27:48 · Dan Lifshits, James Pringle
Because most of what property managers are doing is just conversational project management between the tenant, landlord and contractor on to triangulate communication. So this is a bread and butter for what EA can achieve. And there are many more examples, but I would say that these are one of the most prominent ones where the impact on the user experience is the most felt, to the end customer. [00:28:10] S0: Yeah. Some kind of similarities to some of the technology that previous guest Ali Parsa is using in the health care industry, which is quite cool, that kind of conversational AI element. [00:28:20] S1: I was actually listening to that, and he was finding lots of parallels. It's it's fascinating conversation.
00:28:20 · Dan Lifshits, James Pringle
[00:28:20] S1: I was actually listening to that, and he was finding lots of parallels. It's it's fascinating conversation. [00:28:25] S0: Yeah. It was it was really great. He actually isn't doing the roll up model, but as I'm... As I was hearing you think, speak there, I was thinking, I wonder if he could get up to market a lot faster by buying some companies, like some health businesses, like buying GP practices or something. So, so down, obviously the roll up model is it adds a whole nother level of complexity. You're not just building a product, you're building an M and A organisation as well. So how does that affect the skills required from the founding team and who on the team has taken the lead role on the M and A side? And how has that learning curve been and how many acquisitions have you actually done? Sorry. There's loads of questions there, but Yeah. Just keep No problem. I'm sure there'll be other people thinking about roll ups. And so it'd be great to hear how that all comes together.
00:29:13 · Dan Lifshits
[00:29:13] S1: I think I would agree with you that on one end, we've been talking about that it's effectively almost like buying the customers in a bulk Reality is it's obviously a bit more complex than that because, yes, it comes together with some preserving the existing teams and existing brands and existing processes that you then need to change. And sometimes building from scratch is easier than changing what already exists, as you obviously guys know. So I I agree. I think that this is not for sure an easy business model that, in my view, requires at least three core skill sets around which you're building the business. I would say that first and foremost, you need to have a really, really strong operator that understands how to run a tech enabled version of the service of the business of such. And regardless of effectively how much more efficient can you get with what you're running, it's still always gonna be quite a labor intensive business, meaning that it always gonna be running a huge operational business, number one. So you need to be a very strong, diligent operator first and foremost. That's where I believe that all the experiences with, like, ride hailing and some others comes very nicely together. Second, you're right that you need to have a very strong tech first transformation mindset, the likes of kind of volunteer, forward deployed engineer type, changing the processes and adopting the existing service processes to exchange it into software. So you either need to have a very strong technical person that understands the applied AI, very practical implications of that. And it's sometimes different people from the top AI researchers working on the, let's say, fundamental models itself, because application layer and the fundamental layer are slightly two different things. So we need to understand having the existing operational process, how can I reengineer that with the most progressed existing technologies? So this is number two. And it can be very strong product person, can be very strong technological person. It's amazing if you have someone that brings in some experience from that industry straight away, because then you know already what you are solving and how you're improving that and you know the kind of incumbent status quo.