Mike Grondahl : Got it. And what did you pay for it in 2019?
Orestes Fintiklis: So this is — it’s a difficult question to answer. There was a big share component. But in 2019, we bought a number of business units and a number of businesses. So we basically bought three different businesses. The first one was the hotel part. You may recall, in 2019, we were only selling air. So basically, during the pandemic, we used the content of a business acquired together with this transaction to basically move into the adjacent hotel space, which was very important to our strategy, that business we are not disposing. Then there was a second part, another business unit within the same transaction that was call centers used for servicing customers. That business, again, we are keeping. What we are disposing is a very specific business unit, which relates to selling directly to the consumer, B2C travel through leads generated online and completed through call centers.
So this is a segment of the market, which is really — which is on-site pressure because on the one hand, there is inflation and the biggest inflation is in personnel and human resources. So this is a very labor-intensive business with a focus on selling through call centers, right? So you have these dynamics on the one hand. And then on the other hand, you have suppliers being much, more strict in the trends and the way they are looking at this part of the business, which creates also an incompatibility in our supplier relationships. So this is basically the unit that we have disclosed as part of this divesture.
Jim Dullum: Yeah. And also to add on that, Mike, when you think about the consideration that we got for the sale of this business unit, and the reduction that we had on our intangibles and goodwill, we actually got a profit, right? The cost that you see here is more linked to the agreement on financing and supporting during the transition with the net working capital of this business unit.
Mike Grondahl: Okay. Okay. Thanks.
Jim Dullum: Sure.
Operator: [Operator Instructions] Our next question is a follow-up question from Darren Aftahi from ROTH MKM. Please go ahead. Your line is now open.
Darren Aftahi: Yeah. Darren, I have a broad question about just AI in general, as it relates to your business and sort of with Purplegrids. So there’s, a lot of development costs in training these models. So I’m curious on two things. One, in your own organic marketplace, have you spent sort of the R&D dollars to kind of get that trending up and running, if not, kind of when is that going to hit the P&L? And then for Purplegrids on their customer experience platform, has that R&D investment been made for the most part in the company? And is the product sort of fully up and running? And then as a subsequent question to that, will the CX platform help offset any cost for kind of call centers that you guys may incur today. Said another way, will this benefit your P&L while maybe adding some more costs here and there. Thanks.
Jim Dullum: Thank you for that. I’ll take a first stab and then Prasad will add. So the first part is that we have been already training the models for a long period of time, and they are included in our CapEx. Now a big part of that trend is automatic, because we have already linked our AI. So the bigger part of the cost is developing and integrating it with the operations. And then over time, it’s a declining cost to the actual training. And again, here, we’re benefiting from two dynamics. First, we have a lot of data, 50 million searches are done on our platform every day, which are basically integrated to the AI that keeps learning. And then on top of that, we are training effectively the AI through connecting it with 65,000 experts, so this is a more of direct connectivity in the sense that a consumer is likely to give less feedback and interaction with the platform than an expert by definition has a two-way exchange of information with the provider of the service.
So that’s on your first question.
Prasad Gundumogula: So on the second question, the impact of this platform to our operating business and the efficiencies in our operations, so we have big plans for not only applying this AI platform and our overall AI businesses on call center business, and but also in every part of the business. So we are transforming, we as a Mondee, we want to transform to an AI business inside and outside. So outside that we are giving the information with the models that we developed and the creation that we are planning to work on and with Abhi as our front and foremost application, but also on the back side that we are expecting to transform our current business in every aspect of it, including the call centers, how it actually takes the benefit of this AI driven platforms and as well as the marketing and operations and many other aspects of the business.
For example, our operations, if there’s any — I’ll give you an example, if there is a small piece of cost of sales that we have it is with the debit, memos, charge backs and all. So with this AI platform, we wanted to predict before it happens, throughout analyzing the whole funnel of the transactions and be able to stop it and be able to improve it for the future transactions is a big level that we are currently doing and be able to save a lot of cost of sales dollars.
Jim Dullum: Darren, it’s Jim. Just a follow-up as well. Just to summarize it. Basically on the front end, so for the consumer interface and things like that, a lot of that investment already made. There’s obviously ongoing enhancement and tuning, but a lot of that part of the investments already made relative to putting it across the entire infrastructure. We’re expecting the cost savings to cover most of any further and ongoing investment there. So we won’t quite use the terminology neutral, but I think that the returns from implementation of this particular example Prasad just gave will support most of that investment.
