Mar 4, 2026 Internal connect - AI costing - Transcript 00:00:00 Naveen Puttagunta: um um insurance uh company, insurance brokers database and all of that. There are about 200 uh users on that system, constant data updates. The database is pretty big. um you know it's a regular transactional system not an analytical a trans but it's a Yeshwanth Reddy Yerraguntla: Right. Naveen Puttagunta: fully loaded system in the sense load balancers you know EC2 environments multiple pods kubernetes cluster you know multiple services we use from AWS firewall network everything so the production environment costs us about $1500 per month and then we have we Yeshwanth Reddy Yerraguntla: Okay. Naveen Puttagunta: obviously for IRM we run four environments we We have a production, we have a uh prepro, we have a UAT, uh we have a QA. So four environments together is about $4,000. Okay. Yeshwanth Reddy Yerraguntla: Okay. Naveen Puttagunta: Transactional volume wise, it's not like thousands of users, it is like dozens of users. But it will be true of any mid-market manufacturing company that they're going to ask you about. All right. Yeshwanth Reddy Yerraguntla: All right. Naveen Puttagunta: So that is a typical level of uh infra $1,500 for the 00:01:21 Yeshwanth Reddy Yerraguntla: Got it. Naveen Puttagunta: production system. Of course, you know, if you are bringing like a heavy heavy transaction volume system, obviously it's going to be like something like $2,500. Diya Elizabeth: What do we mean by heavy transaction Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: system? Naveen Puttagunta: uh like hundreds of users um you know continuously Diya Elizabeth: Oh. Naveen Puttagunta: working with the database uh continuously working with application application continuously doing lots of inserts updates things like that okay but Diya Elizabeth: Look. Naveen Puttagunta: the the you know the the infrastructure cost is between that $1,500 to $2,500 Diya Elizabeth: Okay. For the medium size like for medium size Naveen Puttagunta: Mediumsiz transaction systems, Diya Elizabeth: usage. Naveen Puttagunta: it'll be $1,500. For small transactional systems, Diya Elizabeth: Okay. Naveen Puttagunta: it will be like much less. It'll be like $500, $600. Diya Elizabeth: Okay. So if a use nin and like I'll give you a use case that Vijay gave us yesterday. Naveen Puttagunta: What? Diya Elizabeth: He said every he said like right now connecting to those systems may be a problem. 00:02:33 Diya Elizabeth: Right? So one of the things he said is give him an option of you don't even connect to the system yet. We can do that phase two phase three. Every day you're uploading a Excel sheet. Now this Excel has got like 10,000 rows and like 20 columns and every day you're uploading an Excel sheet. Every day you're asking it one question. Naveen Puttagunta: only one Diya Elizabeth: You've got you said one specific question. Naveen Puttagunta: question. Diya Elizabeth: Okay. Say maybe two three questions. Basically from that thing there's one from all of that data there's something that you want to figure out that you're asking every day. Naveen Puttagunta: So let me uh let me reset the thing. Diya Elizabeth: How do you Naveen Puttagunta: That's a very small system. Okay. Diya Elizabeth: like? Naveen Puttagunta: 10,000 rows, you know, 20 columns. It's a very small system. Okay. Um that infrastructure wise because it is an application you have to pro protect it. You know there's a minimum level of infrastructure that you need to spend on okay so that'll be like 00:03:30 Diya Elizabeth: Okay. Naveen Puttagunta: $400 $500 per month. Okay? Because no matter what there's a minimum level of spend right you have to take an EC2 machine you have to put some firewall or things like that you have to do. You can't get away from it because it's a tiny system. Whatever Vijay is talking about is a tiny system. Diya Elizabeth: Yeah. Naveen Puttagunta: Okay. Now I'm going to I haven't even come to the AI costs. Diya Elizabeth: Yeah. Naveen Puttagunta: Um Yeshu I hope you're able to see my screen. This is the analysis I just got pulled from our current uh enterprise brain system. This is the usage in the last 30 days. It is the number of total tokens used, Yeshwanth Reddy Yerraguntla: Okay. Naveen Puttagunta: number of requests and the actual spend in dollars. So per 100 requests, it's about uh 61 cents. Yeshwanth Reddy Yerraguntla: Yeah. Naveen Puttagunta: 100 requests means I I interpreted 100 requests as like one user uh about you know if they're you know if they're doing quite good decent type of analysis on um on the system they might ask about 20 to 25 questions they won't ask more than that in an 00:04:31 Yeshwanth Reddy Yerraguntla: Correct. Naveen Puttagunta: hour because you have to consume the output. Yeshwanth Reddy Yerraguntla: Correct. Naveen Puttagunta: So that 20 25 questions might result in 100 Yeshwanth Reddy Yerraguntla: Yes. Naveen Puttagunta: requests that that ratio makes sense to you. Yeshwanth Reddy Yerraguntla: Got it. Yeah. Fair enough. Yeah. Yeah. Naveen Puttagunta: So about for 100 requests it's about 60 cents. But of course you know uh these these guys are asking simple questions but assuming that the customer will ask very very complicated questions and you have to really you know the number of requests is going to be more or complexity of the request more so the token usage will be higher. So I'm I'm guessing about a dollar. Yeshwanth Reddy Yerraguntla: Yeah. Naveen Puttagunta: So the if one person sits in front of the enterprise brain and asks you know spends Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: H Naveen Puttagunta: like about an hour doing analysis they will spend about a dollar in one one session in one Diya Elizabeth: in one Naveen Puttagunta: day now uh you know like let's say we get to that that's the 00:05:20 Diya Elizabeth: day. Naveen Puttagunta: simplest for the simplest uh system that Simha asked I'm saying in one day means that for one hour session If the guy is continuously banging on the Diya Elizabeth: Correct. Naveen Puttagunta: system for 8 hours, they'll spend eight $8. But the infrastructure cost is the $500 cost like that. Diya Elizabeth: Correct. Naveen Puttagunta: That's the baseline cost. Even 500 is like an overkill for whatever you just described. Okay. Diya Elizabeth: Yeah. Naveen Puttagunta: But I'm just saying because it's a production system, right? There's a minimum level of firewall. Uh you know, all of these things that you need to do. Yes. Otherwise, you can, you know, it's actually less than $100 for the for the use case that Simma talked Yeshwanth Reddy Yerraguntla: Yeah. Naveen Puttagunta: about. The monthly cost is less than $100, right? Yeshwanth Reddy Yerraguntla: This uh uh whatever you're showing on screen is Naveen Puttagunta: Yes. Yeshwanth Reddy Yerraguntla: like this is from enterprisebrain.deom.com the.com is Naveen Puttagunta: Yes. Not enterprise. Yeshwanth Reddy Yerraguntla: it Naveen Puttagunta: I mean this is the token and such from the enterprise brain uh team like you know there I think in light LM there are like multiple team concepts right so from the enterprise brain team I've 00:06:41 Yeshwanth Reddy Yerraguntla: right? Naveen Puttagunta: gotten this data filtered so for the last 30 days that many number of tokens uh this many number of requests and the actual spend was for Yeshwanth Reddy Yerraguntla: Mhm. Got it. Naveen Puttagunta: $43 Yeshwanth Reddy Yerraguntla: Yeah, fair enough. Like this is how Roughly it Naveen Puttagunta: there you Yeshwanth Reddy Yerraguntla: works. Naveen Puttagunta: that Diya Elizabeth: I got till the uh 1 hour usage a day is Naveen Puttagunta: then again don't add 1 hour usage a day. Diya Elizabeth: $1. Yeshwanth Reddy Yerraguntla: Yes. Naveen Puttagunta: One hour usage is $1. Just stick with that, right? Diya Elizabeth: Okay. Naveen Puttagunta: One hour usage for one user is $1. Diya Elizabeth: Correct. Naveen Puttagunta: Yeah. For a simplest, you know, for a simple single source system like what VeJ is talking Diya Elizabeth: Yeah. Naveen Puttagunta: about. Diya Elizabeth: H But is it still a single source system if you keep every day uploading a new data Naveen Puttagunta: Yeah, it's still a single source Diya Elizabeth: set? Naveen Puttagunta: system. Diya Elizabeth: Okay. Okay. Naveen Puttagunta: Okay. 00:07:41 Naveen Puttagunta: Now, um our enterprise brain vision is like dynamic UI. Then every hour we'll be automatically processing data, figuring out insights, all of that. So, the cost is going to go up. Diya Elizabeth: H. Naveen Puttagunta: So in in in our full enterprise brain system that has multiple sources and you do quite a bit of correlation and all of that let us assume that that cost will jump for for one hour of usage the cost will jump to $2. Okay. Diya Elizabeth: Okay. Naveen Puttagunta: So you can say that for the simplest of use cases the cost is $1. For more complex use cases with multiple data sources and all of that the cost per hour is $2. Okay. Then um uh uh and also because I'm doing continuous data refreshes and I'm processing insights and all of that there will be like a you Diya Elizabeth: Hey. Naveen Puttagunta: know whether or not a user uses it there will be a cost because you know in the full enterprise brain system not for the use case that uh Simha just asked and that might 00:08:49 Diya Elizabeth: What am I said? Naveen Puttagunta: be something around uh you know like $20 a day. I mean that's a huge uh LLM usage right um Yeshwanth Reddy Yerraguntla: Yeah. Per Naveen Puttagunta: yes not per user see the in the full enterprise Yeshwanth Reddy Yerraguntla: user. Naveen Puttagunta: brain vision what are we talking about there's an agent that will automatically pull the data then it'll basically run insights it's not user based right it's not per user based it's common it's common to the Yeshwanth Reddy Yerraguntla: Oh, can I do? Yeah. Naveen Puttagunta: system it is going to drive generate insights then it is going to correlate insights it's going to generate escalations, it's going to generate notifications, all of that work it is doing. Yeshwanth Reddy Yerraguntla: Yeah, $20 is like the upper Naveen Puttagunta: So yeah, $20 per day, you know, if in the in the big uh enterprise brain system, Yeshwanth Reddy Yerraguntla: limit. Naveen Puttagunta: it'll probably, you know, use about $20. I I don't want you to give me just that short because then I know I don't I can't figure out whether you're really understanding or whether you're just moving on. 00:09:57 Diya Elizabeth: Understood. Naveen Puttagunta: Okay. So there are three metrics I'm giving you. Number one, there is infrastructure cost which is for the simplest of use cases, it's anywhere between$100 to $300. For a uh midsize, you know, like enterprise uh application, the infrastructure cost is about uh $1,500. That's a quite, you know, like 200 users uh touching the system every day. Lots of transactions, heavy heavy heavy transactional usage will be like $2,500 to $3,000. Um for the base infrastructure cost for the application for the use case that Simha mentioned, it is as low as probably $100 $200. Okay. Then that is the infrastructure bucket. Diya Elizabeth: Check. Naveen Puttagunta: Then the LLM usage bucket is two separate things. One is when a user is interacting with the system and they're asking questions and getting answers and all of that. So they're doing analysis. uh for per hour of usage per hour of analysis it's about $1 max of $2 like even full Diya Elizabeth: Okay. Naveen Puttagunta: dynamic UI and all of that will also be $2 up to $2 per hour of analysis. 00:11:19 Diya Elizabeth: Okay. Naveen Puttagunta: Then the third bucket is oh you know in you know if we have automatic data processing and we're retrieving data and and doing you know figuring out insights that's not the use case that Simha has this is like our vision of enterprise brain multiple data sources correlation all of that if you do all that kind of stuff then per day it's about uh you know it might be about uh $10 to $20 uh per day usage. Diya Elizabeth: cost for implementation that's much uploading Excel sheets and reading from Excel sheets is how easy is it and what would be the cost of implementation okay per month you remember that at home you first tell Then I will ask you the Naveen Puttagunta: that. Diya Elizabeth: second third Naveen Puttagunta: So, those are the three buckets that I gave you. Now, Diya Elizabeth: bucket. How did you say Nin? Sorry, I missed the number with myself. Naveen Puttagunta: uh $10 to $20 per day. that is not user based. That is you know the system is automatically pulling data, 00:12:19 Diya Elizabeth: Okay. Naveen Puttagunta: refreshing data, processing, generating insights, generating uh escalations, generating notifications, all of that. That is not the use case that Sima asked. Diya Elizabeth: H. Naveen Puttagunta: This is our enterprise brain vision. Diya Elizabeth: Yeah. Naveen Puttagunta: Now, Diya Elizabeth: Yeah. Naveen Puttagunta: next question is you were asking cost of implementation. Before I go to cost of implementation, um there was this oh you know if you're doing um um data science based statistical modeling uh you're running predictive analytics you're running classification problems you're running regression problems that kind of stuff then again it it comes under infrastructure cost it is not LLM based that is infrastructure cost it is uh if if you if you're using that uh GPU PU Diya Elizabeth: Okay. Naveen Puttagunta: heavily and then the uh generally the GPU machine costs are higher. So rather than that uh you know so you add on like $200 to uh your infrastructure cost per Diya Elizabeth: Okay. Naveen Puttagunta: month Diya Elizabeth: So what Simma said Naven is that then is it which one would it be using? Is it stat statistical modeling or is 00:13:42 Naveen Puttagunta: no most likely. Diya Elizabeth: that Naveen Puttagunta: So I because it's not clear what he's asking because it's a very generic statement. Diya Elizabeth: h Naveen Puttagunta: See if it is statistical modeling and all of that then this LLM cost will not be there. Okay. Diya Elizabeth: understood? Naveen Puttagunta: The infrastructure cost might slightly go up. Okay. For SIMHA's use case because the first of all the application is so small the infrastructure cost is Diya Elizabeth: Okay. Naveen Puttagunta: just just an add-on. So it will not be significant the you know that GPU cost will not be significant. Diya Elizabeth: Okay. Naveen Puttagunta: Okay but uh for L if it is an LLM based then it is the you know the uh the it's a small system and the LLM usage is like a dollar per day at a dollar per session per per hour of analysis dollar per hour of Diya Elizabeth: for one use. Naveen Puttagunta: analysis. Okay. Now cost of implementation for something like that I Diya Elizabeth: Yeah. Naveen Puttagunta: would say uh it is maybe uh $20,000 you know if you if you want to do a proof of concept um I could do it in like 15 $20,000 maybe even less simply taking in that uh uh spreadsheet uh dumping that data into database and running enterprise brain on top of that. 00:15:10 Naveen Puttagunta: Yeah, you know, we can say that, you know, especially if you're trying to land a deal as a P, we could do it under $20,000. Diya Elizabeth: Yeah. Naveen Puttagunta: Probably even less. Diya Elizabeth: One minute. Naveen Puttagunta: I mean, the true cost is probably much lesser. Yeah. I mean, what is the baseline cost? I would say maybe I would say that is like really trivial. So if we had the rest of enterprise brain, I'm not char if I'm not charging for the rest of enterprise brain, then it is really just about uh I can do it under $5,000, but I am also going to charge for a little bit for enterprise brain, right? I mean, I'm not giving the enterprise brain for free. So that is why I would bump it up to let's say $20,000. You this part you discussed with Pratima because he says don't give numbers, Diya Elizabeth: Yeah, that's what Naveen Puttagunta: right? Diya Elizabeth: I Naveen Puttagunta: I'm saying the the use case that Simha talked about if it is really that trivial we could do it within one person month one person working for one month we can deliver that if I'm not assuming that 00:16:18 Diya Elizabeth: Yeah. Naveen Puttagunta: the rest of enterprise brain is already there I'm not charging for the end then on top of them then you have to charge for the enterprise brain system you know because that is already developed that to add Diya Elizabeth: Did it. Naveen Puttagunta: on so that you can do whatever you want. Diya Elizabeth: Okay. Naveen Puttagunta: This I feel pratima needs to be involved in this specific Diya Elizabeth: Right. Naveen Puttagunta: discussion. Diya Elizabeth: I was just We can't discuss. We have to go. Come. Okay. It's uh Naven, we have a meeting in half an hour. Naveen Puttagunta: Yeah, Diya Elizabeth: We have to leave Naveen Puttagunta: I understand. I just only one two line sentences. Diya Elizabeth: now. Naveen Puttagunta: Pma just hear these two sentences that's all. Okay. Diya Elizabeth: Okay. Naveen Puttagunta: See to cost of development pratima for the very specific use case that uh Vijay mentioned it's a very trivial it's it's less than one person month of effort I would say that so therefore it's like $5,000 uh but assuming that the rest of enterprise brain is already there because rest of 00:17:24 Diya Elizabeth: One second. VMR's use case where you have to upload one sheet like the Excel sheet process that Excel sheet just that the implementation not the in not the LLM not all of that just the implementation cost is $5,000 implementation means uh giving them the flexibility to be able to upload Excel sheets over a period of time that only to do that connector it is Naveen Puttagunta: Yes. Diya Elizabeth: $5,000 Naveen Puttagunta: And and process it and put it into the database. Right. Assuming that the rest of enterprise brain is already Diya Elizabeth: okay uh Yeah. Naveen Puttagunta: there. Diya Elizabeth: So, enterprise brain, right? We also I get it now in the cost of implement cost of enterprise brain is not there. I told him $1,000 because you know the way we when we have been talking it's just an Excel sheet is how I thought. Did you tell the number of rows? Yeah. 10,000 or Huh? Naveen Puttagunta: They don't matter. Pratma number of rows, Diya Elizabeth: They don't matter. Naveen Puttagunta: number of columns, 00:18:21 Diya Elizabeth: Okay. Naveen Puttagunta: they don't matter. Diya Elizabeth: Okay, because it was such a small Excel sheet, we thought uh okay fine now implement it. Naveen Puttagunta: No. Diya Elizabeth: I mean see for such things if I ask them uh pay for the entire enterprise brand they won't pay kadavin. So how do I do that is one thing that we have to figure out Naveen Puttagunta: you I mean so you have to Diya Elizabeth: while Naveen Puttagunta: decide how would you amortize the cost of enterprise brain right it's a one-time development cost for us it is probably about 2C cr enterprise brain the whole thing is about 2C cr for us so how will you amortize it you have to figure out Okay. Diya Elizabeth: Oh, Naveen Puttagunta: So amort by the way the whole enterprise brand I would say it's about 2 to three CR PMA Diya Elizabeth: which would be how many dollars? Naveen Puttagunta: uh 350 $400,000 but obviously you can't charge that on one customer. You have to figure out how to you know get value out of it from multiple instances. Diya Elizabeth: How many get how many customers if we get your enterprise brain will like the enterprise brain will be profitable if I were to charge right for discovery right I'm thinking on an average of 20k and use two use cases two simple use cases per use case 10,000 I mean I'm just I'm just throwing it out there you will need one 5 00:19:54 Naveen Puttagunta: See Diya Elizabeth: 10 15 users to even get the cost of Naveen Puttagunta: no see for with every engagement you spend additional amount Diya Elizabeth: enter Naveen Puttagunta: right because you have to connect that particular data set you have to try some some things out. You have to configure some things. You have to teach the system some things. So imagine that for every engagement you are spending at least uh four weeks of um like one person months worth of effort. Some customization will be required. So that one person customer one person uh effort is assume that that is uh uh $5,000. Okay. So if you're charging $20,000 you're recouping $15,000 towards enterprise bid. So if you're doing $15,000 into let's say 30 customers, you're going to recoup the cost of enterprise brand. Diya Elizabeth: I miss you. Naveen Puttagunta: That's Diya Elizabeth: Got it. Naveen Puttagunta: it. Diya Elizabeth: Maybe to begin with to get in, we may have to charge less and over a period of time we can increase the cost. Maybe we we'll figure it out. 00:21:04 Naveen Puttagunta: You figure that out. Now that you understand the cost, Diya Elizabeth: But let me get Naveen Puttagunta: you figure it out. Yeshwanth Reddy Yerraguntla: Yeah. Naveen Puttagunta: Yes. You you have everything right. Diya Elizabeth: okay. Yeshwanth Reddy Yerraguntla: Yeah, I Naveen Puttagunta: If there is any questions you you can Yeshwanth Reddy Yerraguntla: do. Naveen Puttagunta: answer. Diya Elizabeth: uh you uh we this meeting is at 10:00 mark so it's more of a Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: discovery so I need you to participate to ask all the technical Yeshwanth Reddy Yerraguntla: Mhm. Diya Elizabeth: questions okay because I don't know I mean I I almost spent about 3 hours this morning just trying to understand their business um what they do and all of that uh I have some decent understanding uh so while I'll be driving the conversation maybe. Let's see. We will have to figure out other things. Okay. Yeshwanth Reddy Yerraguntla: Uh what is the company? Diya Elizabeth: Same trace yesterday Yeshwanth Reddy Yerraguntla: Trace Raycast Diya Elizabeth: evening. Yeshwanth Reddy Yerraguntla: like Diya Elizabeth: Okay. Bye. 00:22:09 Diya Elizabeth: We have to leave. Yeshwanth Reddy Yerraguntla: Yeah. Naveen Puttagunta: Okay. Diya Elizabeth: You have any other questions? Naveen Puttagunta: I have one question. Yes. when they ask for Okay, Diya Elizabeth: I otherwise I won't be able to eat breakfast and I want to eat. Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: I'm hungry. Naveen Puttagunta: then why don't you just take the move this to the phone? Diya Elizabeth: I I I will figure that out. You continue talking. Naveen Puttagunta: No, no. I needed you and Pratima to be there. That's Diya Elizabeth: 2 minutes we have to get ready and then uh take Naveen Puttagunta: why Yeshwanth Reddy Yerraguntla: What? Diya Elizabeth: over when we are packing right when we are taking meeting and packing we'll forget a lot of things. So I need a piece of about 2 3 minutes. So, Naveen Puttagunta: Okay, Diya Elizabeth: and then at Naveen Puttagunta: only one thing how you talk to Yashwan on how he should answer. Diya Elizabeth: the Naveen Puttagunta: Oh, how hard is it? How how long will you take? 00:22:55 Naveen Puttagunta: Those kinds of questions because they will then tie that to your pricing later. So you figure that out what Diya Elizabeth: Yeah. Yes. Yes. Naveen Puttagunta: should Diya Elizabeth: One thing, one thing while I really like the way you presented that, you know, we don't have to redo everything yesterday. Remember, I don't know which meeting you said that in those two meetings. I also want you to rephrase certain things because if we don't have to while the technology is common across then why are you charging me so much and understood? Yeshwanth Reddy Yerraguntla: Understood. Diya Elizabeth: So while you meant acceleration and I mean and uh don't worry we know Yeshwanth Reddy Yerraguntla: Yes, Diya Elizabeth: technology subject matter expertise absolutely I'm with you on it. Yeshwanth Reddy Yerraguntla: that's what I Diya Elizabeth: The downside is okay if you the technology is the same if all of this is then why are you charging me so much Yeshwanth Reddy Yerraguntla: meant. Diya Elizabeth: any soar I have to figure out how to present but I'm absolutely with you on that point that the points are absolutely right and stick to those points but how to present we figure 00:23:56 Yeshwanth Reddy Yerraguntla: Understood. Yeah. Diya Elizabeth: it out you understood my point right because I do want to tell them that underlying Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: technology we understand you give subject matter expertise we know how to build the AI so you don't worry type say I want to And it is accelerated also. I want to Naveen Puttagunta: No matter what, don't hesitate answering, right? Don't hesitate answering. Yes, we can do it. Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: tell. Naveen Puttagunta: Kind of because when you hesitate, right, people think that, oh, these guys can't actually do it. They're just fudging. So, just make sure that the vibe is that, hey, okay, this is I understand this technology. No problem. We can build it. Yeshwanth Reddy Yerraguntla: Yeah. Diya Elizabeth: Yeah. Naveen Puttagunta: How long will it take and all of that? Diya Elizabeth: In fact, Naveen Puttagunta: You you figure out what is a strategy to answer with the Yeshwanth Reddy Yerraguntla: Got Diya Elizabeth: no. In fact, Naveen Puttagunta: pratma. Diya Elizabeth: Yashwant was doing good yesterday. I mean both DIA and Ashwant we we we move to the next level of sales so I'm Yeshwanth Reddy Yerraguntla: it. Naveen Puttagunta: Fantastic. I want Diya Elizabeth: happy I don't know sometimes I'm just leaving it to her and let her Naveen Puttagunta: Yes. Diya Elizabeth: do and then I'm interfering when required and so did Yashanti yesterday so the way he presented I really liked it and I really like that voice Yeshwanth Reddy Yerraguntla: Thank you. Diya Elizabeth: thing come up because then there people are realizing it's voice the first meeting you didn't have Yeshwanth Reddy Yerraguntla: Yeah. Yeah. Naveen Puttagunta: I don't know what thing is but I would like to know more but later. Yeshwanth Reddy Yerraguntla: We'll have to figure that. Naveen Puttagunta: Okay, bye guys. Yeshwanth Reddy Yerraguntla: Bye. Naveen Puttagunta: Good luck on your meetings. Yeshwanth Reddy Yerraguntla: Thank you. Naveen Puttagunta: Have a blast and then um close deals. Bye. Yeshwanth Reddy Yerraguntla: Bye. Transcription ended after 00:25:31 This editable transcript was computer generated and might contain errors. People can also change the text after it was created.