Feb 25, 2026 Meeting Feb 25, 2026 at 18:58 PST - Transcript 00:00:00 Yeshwanth Reddy Yerraguntla: architecture whatever steel we presented something and they were very happy with it right so detailed enterprise brain we are this is like we this is our current understanding of how it should be dep so that all the okay extensibility and all the features that we want also come into ure definitely it's easier to it's easier and faster to implement this rather than retrofit that so Navin Garu what he suggested is I commit to you how many hour days uh is needed to come up with the right uh set of backlog the output of planning is uh backlog How we will start, when we'll start, that's not the point. But to get to to get this thing uh on code or in code and working, how whatever tasks are needed uh put them uh in the form of a backlog and uh then we can worry about how to prioritize it, whom to give each task to, how to estimate everything. which which was the missing step. Whatever I was trying to say, I was unable to convey. 00:02:05 Yeshwanth Reddy Yerraguntla: But yeah, this helped me a lot. Satyasri Prabhakar Mantripragada: Fine. We all have to brainstorm together to understand what do we want to accomplish? What was the vision? Where do we stand? Yeshwanth Reddy Yerraguntla: Yeah. Good. Satyasri Prabhakar Mantripragada: How long would it take to finish that? Yeshwanth Reddy Yerraguntla: Yeah. Satyasri Prabhakar Mantripragada: We are jumping into it and without even a plan or without a high level Yeshwanth Reddy Yerraguntla: Yeah. Satyasri Prabhakar Mantripragada: okay we don't have the Yeshwanth Reddy Yerraguntla: Oh, Satyasri Prabhakar Mantripragada: right skill or something something lay Yeshwanth Reddy Yerraguntla: got it. Satyasri Prabhakar Mantripragada: down okay how do we want to go what do we need whom do we need to train if needed then the second thing second layer is the time Current map is Yeshwanth Reddy Yerraguntla: H yeah yeah yeah that will Satyasri Prabhakar Mantripragada: record. Yeshwanth Reddy Yerraguntla: come anyway after everything on is available on paper you only commit to the planning part of it right now at the same time let the arrest of the people continue working because there will be some components which are critical 00:03:20 Satyasri Prabhakar Mantripragada: Huh? They'll continue. Yeshwanth Reddy Yerraguntla: anywhere. Satyasri Prabhakar Mantripragada: Got it. Yeshwanth Reddy Yerraguntla: So what I want to propose is for this uh Satyasri Prabhakar Mantripragada: Yeah. Yeshwanth Reddy Yerraguntla: I want to take your help your time Manisha's time raw's time uh Kud Pawan and Ganesh time also because they are seniors in uh understanding the architecture uh for whatever we want to build. First we have to understand what is this. Second thing will be inside each one of this uh at a very high level what we are saying is okay UI architecture only core architecture okay data source architecture data source is basically all the individual agents right Google suit Google mail agent Google chat agent Google docs agent Jira agent uh Salesforce agent like that What is happening in each of these orange boxes? We have tried to expand on it. Nonetheless, those uh these orange boxes are talking with the red box. Red box is the core agent. Core agent itself has two different uh um AI components. 00:04:38 Yeshwanth Reddy Yerraguntla: Okay. it one that is necess uh that is interfacing with the uh data source agents and one that has to interface with the UI part of it because uh this agent uh it is its job is to mainly analyze what is the state of my organization currently right because every individual agent will say okay uh this is I fetched the mails and I'm ready to give give you answers to questions based on whenever you are asking. This is a reactive uh uh state of uh functionality. This they are just uh sitting there waiting for someone to ask a question and when you ask I will give you answer to. So who is asking is this red box and this red box is also going to ask based on the state of the uh uh organization. Like if agent this agent knows that currently these are the seven projects that are happening then it makes sense to ask uh only questions about those seven and updates it update its memory of what is happening in those seven projects right because five Jira task one of them actually regressed I received eight to 10 emails from different different clients in this uh whatever uh Google chat um These are the rough updates. 00:06:11 Yeshwanth Reddy Yerraguntla: One person is on leave. So that is all knowledge for that one project. Like that it has to keep updating its own understanding of what the seven projects are. At the same time it will continuously ask hey is there an eighth project or a ninth project in the pipeline. Salforce right. So it know it needs to know what it should ask and whom it should ask. On top of that, you have an agent which is deciding what to show to every person. If if I have an snapshot of my seven projects, not everything in every detail needs to be uh explained to every person in the organization, right? So, Navinana will probably only want to understand is everything sailing smooth or not. Pratima madam will be interested only in is customer uh uh uh happiness maintained or not? Customer satisfaction you as a uh uh delivery uh lead right you are interested in are there any uh uh like uh is the burn rate proper or not uh for you? 00:07:27 Yeshwanth Reddy Yerraguntla: Yeah. So every person's slice of that project is different. Am I making sense? Rajashekar G: Yes or Yeshwanth Reddy Yerraguntla: Yeah. Like if there are 20 updates, Rajashekar G: no? Yeshwanth Reddy Yerraguntla: maybe only one update needs to go to Navin and that update could be a summary of five rows. Uh one update needs to go to Prabhakar. That is a summary of eight rows but a different eight rows. So that is handling by the user aware alert distribution agent. Visualization intelligence is like okay now that I received these inputs what are the right ways to show them? Should I should I use bar charts or should I use something else? Whatever. There is Rajashekar G: I have one question here just I want to uh make sure like core Yeshwanth Reddy Yerraguntla: ask Rajashekar G: agent part low. So we are saying there will be overall three agents or two agents Yeshwanth Reddy Yerraguntla: two agents. Why did you think third will be there? 00:08:29 Yeshwanth Reddy Yerraguntla: Refresh Rajashekar G: uh in the uh refresh agent is there in the bottom expansion Yeshwanth Reddy Yerraguntla: agent. Rajashekar G: details agent Yeshwanth Reddy Yerraguntla: Yeah, I will I will explain what is this refresh agent. Rajashekar G: key Yeshwanth Reddy Yerraguntla: We just called it refresh agent but uh uh Rajashekar G: something something acts like a schedule basically Yeshwanth Reddy Yerraguntla: it it is nothing but a schedule. Yeah. So agent part comes probably to fine-tune for every data source what is the refresh schedule. For Jira the schedule could be once every 6 hours. For emails it could be once every 10 minutes. For something else it could be one per day. Rajashekar G: Uh will that be a configurable rules engine or we will give instructions through NLP? Yeshwanth Reddy Yerraguntla: We'll have to figure it out. We'll have to figure it out. Rajashekar G: If you do it through NLP that will become an agent again if you doable matter that won't be Yeshwanth Reddy Yerraguntla: Correct. Correct. Exactly. 00:09:20 Rajashekar G: an agent. Yeshwanth Reddy Yerraguntla: We have to discuss that. Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: It might so happen that in the interest of time we'll say we'll just hardcode this table. But once we have some shown some maturity, we'll then introduce an agent which will update that table. Rajashekar G: Yeah. Yeshwanth Reddy Yerraguntla: So we have to make those uh decisions uh uh through and through. But the point is both these items should get into the backlog that eventually yes we want to build an agent and we want to have a way to update this table. Both are two different tasks. Both have different priorities, different timelines, Rajashekar G: Got Yeshwanth Reddy Yerraguntla: right? Unless we put it on paper, Rajashekar G: it. Yeshwanth Reddy Yerraguntla: we will not be in a position to prioritize. That is what I understood from Navincer. All right. Rajashekar G: Yeah. Yeshwanth Reddy Yerraguntla: So the what is the point of this refresh schedule is uh this this enterprise brain what you're building is a very deeply uh proactive uh system. 00:10:23 Yeshwanth Reddy Yerraguntla: It has to tell me something is going wrong. It has to tell me everything is all right. I will not go there and ask. Why? Because simple. If I start asking, it will answer me. Rajashekar G: All Yeshwanth Reddy Yerraguntla: But how many will I ask and what is the guarantee I'm asking the right questions? I don't know like problem because I don't know that Rajashekar G: right. Yeshwanth Reddy Yerraguntla: there is a problem. I don't even know how to ask it. Whereas agent itself what it should do is at every interval it will get a ping from the refresh agent saying okay I think it's time to refresh your knowledge of Jira. It's time to refresh your knowledge of emails. At that point, it will ping the data agents because this said uh ping refresh your Jira knowledge. It will go and ask Jira agent uh please give me the latest updates. Jira will know what to give updates in the form of information alerts whatever it we are defining what Jira thinks should be escalation insight alert information. 00:11:27 Yeshwanth Reddy Yerraguntla: the reactive mode always exists that when I ask a basic Rajashekar G: All Yeshwanth Reddy Yerraguntla: ad is always there that is one small use case in the whole pipeline but the rest of it is uh someone is pinging it saying hey it's 5:00 go check uh if something is updated Rajashekar G: right. Yeshwanth Reddy Yerraguntla: in Jira then this realtime agent will ask Jira agent can you give me something if updates updates in Jira Jira will say these are the 20 updates. Those updates it will go and put it in its uh knowledge graph. Rajashekar G: Uh I have two questions here. Uh just for better clarity. Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: So it's like it is continuous monitoring or uh refresh uh time based uh Yeshwanth Reddy Yerraguntla: Yes, Rajashekar G: fetching. Yeshwanth Reddy Yerraguntla: it is time based. Rajashekar G: Why I'm saying this? Yeshwanth Reddy Yerraguntla: It is polling right now. But yeah, continue. Rajashekar G: Why I'm saying this means uh let's say Salesforce or Jira whatever it may be. So there will be continuous data changes will be happening there. 00:12:32 Rajashekar G: So are we monitoring the data changes in Jira continuously or every 1 hour or 2 hours or whatever the Yeshwanth Reddy Yerraguntla: Correct. Rajashekar G: scheduled time? Are we comparing it with the previous snapshot and current snapshot? What are the changes and different mechanisms Yeshwanth Reddy Yerraguntla: We are doing both. Rajashekar G: complete? Yeshwanth Reddy Yerraguntla: We are doing both. I'll tell you what I mean. Uh first of all in a best case scenario this whatever Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: where is it with this Jira agent would itself would have said hey there is something important go tell it to the right persons correct that ideal Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: scenario the only way Jira would have known something is important is when Jira where is it this Jira data source this this green box itself would ping ping the agent saying something is wrong or something is updated but a ping mechanism data sources lab these are all just APIs so someone has to Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: constantly go and fetch uh constantly keep asking hey is there an update is there an update is there an update because that's the best thing we can do so 00:13:44 Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: periodic uh refresh is inevitable it's we don't I don't think We have a better solution. Unless you guys can have a different Rajashekar G: Uh, Yeshwanth Reddy Yerraguntla: idea Rajashekar G: I mean continuous monitoring periodic refresh Yeshwanth Reddy Yerraguntla: continuous Rajashekar G: point Yeshwanth Reddy Yerraguntla: monitoring Rajashekar G: that is what continuous ping. If something changed in Jira that should directly reflect into Yeshwanth Reddy Yerraguntla: how continuous how continuous is the question, Rajashekar G: our Yeshwanth Reddy Yerraguntla: right? I will ping every 1 second or 10 seconds or 1 minute or 10 Rajashekar G: right got Yeshwanth Reddy Yerraguntla: minutes. Rajashekar G: it Yeshwanth Reddy Yerraguntla: That's I'm saying whatever is the frequency it has to be there. It you put what what is the difference between polling versus uh uh what notification system? Rajashekar G: schedule Yeshwanth Reddy Yerraguntla: Notification is a push mechanism. Here we are saying this is a pull mechanism. Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: I have to constantly pull from sources. Shall I continue? Okay. 00:15:00 Rajashekar G: Yeah. Yeshwanth Reddy Yerraguntla: So I was saying okay Jira said uh there are the these are 10 task updates. It will update its knowledge graph. Agent this red box is responsible for updating this knowledge graph. Now agent will also ping the user aware engine saying uh I have some insights and alerts. U you have to figure out which people to ping the information to. fetch information from user preferences because some users want bar chart, some users want something else. That user response uh what do you call user persona related information is also something that needs to be uh stored and tracked only then we can personalize Rajashekar G: right here Yeshwanth Reddy Yerraguntla: that Rajashekar G: uh the knowledge graph is part of knowledge base. Yeshwanth Reddy Yerraguntla: knowledge graph is yes this is the enter knowledge like this. Rajashekar G: In knowledge base I will be having three things but knowledge graph is one part of knowledge base. Yeshwanth Reddy Yerraguntla: Yeah, Rajashekar G: Yeah. User Yeshwanth Reddy Yerraguntla: you have user related knowledge. 00:16:26 Rajashekar G: information. Yeshwanth Reddy Yerraguntla: You have organization snapshot snapshot show and you have organization level information when there is a bug uh in the system uh in so and so uh someone reported a bug in uh Google chat what needs to be done some people need to be pinged some tickets need to be updated uh steps as an organization some things need to be logged somewhere some people need to be informed that is a Rajashekar G: right? Yeshwanth Reddy Yerraguntla: runbook right um one level below runbooks are skills skills is like how to what are the right ways to ping Jira what are the uh clarity what how to differentiate skills resources around books. It we have to discuss that resources on the other hand is Rajashekar G: Uh I mean runbook say is something comes from the user cover Yeshwanth Reddy Yerraguntla: a Rajashekar G: like if the user means uh let's say there is different role called trainer next time system prompt update uh he will be updating the system prompt again right technically means he's changing the Yeshwanth Reddy Yerraguntla: Yeah. 00:17:56 Rajashekar G: rules but skills are different like skills are like uh Yeshwanth Reddy Yerraguntla: Yes. Rajashekar G: uh let's say salesforce is there what tables are important and what are the major uh tables you need to fetch all the time on every skill Yeshwanth Reddy Yerraguntla: Correct. Perfect. That's pretty much what it should Rajashekar G: But over the period of time if someone comes and Yeshwanth Reddy Yerraguntla: be. Rajashekar G: says Yeshwanth Reddy Yerraguntla: skill skill is also still trainable on because it's specific. Okay, Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: one way to look at it is skill is uh data related or data specific runbook is process specific processes Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: across uh vertical across people across uh uh domains across data sources. Rajashekar G: No. Yeshwanth Reddy Yerraguntla: skill is like specific to one uh data source or one uh uh something more narrow runbooks use skills still uh I will still accept Rajashekar G: Okay. Right. Yeshwanth Reddy Yerraguntla: that this is not a wellought out uh this thing why I'm saying it is also because there are also skills in uh data sources skills on 00:19:23 Rajashekar G: Major data sources. Yeshwanth Reddy Yerraguntla: tech network Rajashekar G: Oh. Yeshwanth Reddy Yerraguntla: knowledge again um we have to we have to sit and discuss what what needs to go and sit in each one of them are the organ information this is the snapshot this is the user info these three are necessary to deliver the right information to the user so this is the architecture of the core agent Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: Again, so we have to understand on technically on on paper how to achieve these things. What does it take? Now here we have we just discussed these two. Oh sorry only this one we discussed the orange boxes I feel is the simplest of all of them. Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: Uh all it is doing is talking only with the real time guy. This is a reactive agent. Uh this has its own understanding of what it is trying to deal with. Uh concepts. I I wrote a bunch of jargon but you have to really define what these mean. 00:20:40 Yeshwanth Reddy Yerraguntla: Few things very clear to me are what does it mean to have an escalation for that data source? What is the snapshot for that day? And what is the world model? uh rag law we are having more uh flexible information like skills and training patterns organization knowledge data resources are static documents and this is this data source this this small pipeline is NLSQL this very small thing this and Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: Anyway, uh finally UI cost UI I I still feel we can do better here. UI has a dashboard. UI has a chat interface. Chat interface is where people ask ad hoc queries. people ask uh and get back answers in the chat. Uh chat interface is where training happens. You do this this this this and no more. And uh coming to dashboard visualization agent will figure out based on user preference and insights coming from the core agent render core agent instances we can c it here itself. 00:22:15 Yeshwanth Reddy Yerraguntla: Now that I think about it, it's based on token amount. Yeah. Yeah. User preferences. It will just put it on the dashboard and say okay this is what Prabhakasar needs to know. This is what needs to know. This is what Rashika needs to know. One alternate uh line is the critical messages. Somehow the core agent will decide okay something is actually very critical. I will bypass the whole thing and I will send it to their email or mobile phone. So the architecture at a very high level said yeah this looks good. Now we have to start understanding how long it is going to take to come with the black backlog. After which um how to prioritize execution. Meanwhile, we also have to ensure our planning and uh uh that people are attending the meetings. Jira board everyone has their tasks full backlog is accepted daytoday people are updating their Jira task we have to freeze them we have to come up with a set of instructions for everyone to 00:23:52 Rajashekar G: Yeah. Yeshwanth Reddy Yerraguntla: Hello. Rajashekar G: And one clarity I needed here the planning backlog means so in the complete new implementation are we going to see what we put it in the backlog and what we prioritize or right now what whatever we have built Yeshwanth Reddy Yerraguntla: This is already present. That is already present. You can suggest at that point we'll add those Rajashekar G: Okay, got it. Yeshwanth Reddy Yerraguntla: notes. Rajashekar G: Got it. Yeshwanth Reddy Yerraguntla: So uh one thing we generally don't take seriously is the testing part of it. Adura I want to um kick off. By the way our deadline is 6 weeks from today 6 and half on coach right uh because March 2nd is the kickoff date that is week one. We are in week zero. Uh week zero Thursday. Um yeah. So at a very high level this I am opening the I just want to now shut up and hear your thoughts. Manisha Gundapuneedi: One confirmation is whatever architecture uh is there on the right side that was approved by Nam or may we have to make 00:25:36 Yeshwanth Reddy Yerraguntla: This was approved by Navenar in the sense that this is looking Manisha Gundapuneedi: changes. Yeshwanth Reddy Yerraguntla: good. We have to make changes on. I just added this thing called skills, right? I did one more change where I said user preferences. Those are the kinds of small small changes we'll realize while doing it on the Manisha Gundapuneedi: But overall architecture wise this is how we are going to go Yeshwanth Reddy Yerraguntla: fly. Correct. Correct. Manisha Gundapuneedi: it. Yeshwanth Reddy Yerraguntla: Correct. End of the uh activity we have four locations of intelligence. Visualization, user awareness or alert distribution, core intelligence and data source intelligence. So people are agreeing. Gopals are also said yeah this looks good features Manisha Gundapuneedi: Okay. Yeshwanth Reddy Yerraguntla: whatever we want to implement. No at a large level everything we want to achieve is being is getting captured in this Rajashekar G: Okay. Now I have one point. So while planning itself right now, 00:26:46 Yeshwanth Reddy Yerraguntla: Hey, Rajashekar G: what are the data sources we are targeting? Clarity only then intelligence means intelligence Yeshwanth Reddy Yerraguntla: hey. Rajashekar G: intent very thin lines the separation full-fledged clarity like what is the real expectation out of the system let's say whatever we have we are doing since last 3 weeks or two weeks Gmail still we are figuring it out Yeshwanth Reddy Yerraguntla: Oh Rajashekar G: But it's Yeshwanth Reddy Yerraguntla: no. Rajashekar G: like Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: again new source comes. First thing we do is getting the right data first. So that might help much Yeshwanth Reddy Yerraguntla: Okay, Rajashekar G: better. Yeshwanth Reddy Yerraguntla: one point I understood is let's uh also try to target the agents themselves which we have to Rajashekar G: Yes. Yeshwanth Reddy Yerraguntla: build huh based on the Rajashekar G: Based on agents the complexity will change. Yeshwanth Reddy Yerraguntla: agent you'll understand um are we covering all the edge cases or am I missing anything I want also no I'm not saying we'll deliver these things we'll think through how to create the architecture. 00:28:41 Yeshwanth Reddy Yerraguntla: You think while thinking these are the things we want to Rajashekar G: G drive should be there. Yeshwanth Reddy Yerraguntla: deliver. Rajashekar G: G drive and uh some custom database like say insurance database. Yeshwanth Reddy Yerraguntla: Okay. Rajashekar G: I think the Yeshwanth Reddy Yerraguntla: Yeah. Sure. Sure. Sure. Yeah, at least for the sake of coming up with the architecture, it makes sense to cover all of these guys as part of the plan. Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: If I'm not wrong, Rajer, what you're saying is this this box needs to be uh created for each of those bo uh seven agents on Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: right individually only then Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: we'll know what are the what is the supererset of all the features. Rajashekar G: Right. Yeshwanth Reddy Yerraguntla: Okay. Rajashekar G: I meant to say now based on that uh data source itself the complexity and planning will change lot of Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: things Yeshwanth Reddy Yerraguntla: Got it. The first thing we have to pro commit to is how long are we going to take for planning and uh 00:30:59 Rajashekar G: All right. Yeshwanth Reddy Yerraguntla: the output of this planning will directly feed into every one of these things. So I want to yeah I want to start with uh okra everyday. Rajashekar G: both ways work and we need to involve other people like Paw and Ganesha common discussion point to feel a little bit Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: better. Yeshwanth Reddy Yerraguntla: Yeah. Only half time from Yeah, we first of all let's agree on a time uh slot when we want to have these Rajashekar G: All Yeshwanth Reddy Yerraguntla: discussions every day. Best case scenario every day we discuss for uh 4 hours. Best case Rajashekar G: right. Yeshwanth Reddy Yerraguntla: 7 sorry 7 in my time is 8:30 in your time and morning and you Rajashekar G: In the morning. Yeshwanth Reddy Yerraguntla: know 10:30 drop off 10:30 is 11:30 12 in your time 8:30 to 12 30 12 3 and a half hours evening uh probably I can review uh if there are any pending items or evening we can connect for half an hour 3 and half hour is sufficient it's okay but are you okay with this 00:33:05 Rajashekar G: Okay. Yeshwanth Reddy Yerraguntla: time office Manisha Gundapuneedi: morning. Rajashekar G: Not Manisha Gundapuneedi: Yeah. Yeshwanth Reddy Yerraguntla: distractions. Manisha Gundapuneedi: Oh, Rajashekar G: one challenge I can see here like morning time discussions regular in the Manisha Gundapuneedi: Um available the graph was soften. Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: morning and sometimes late evening starting starting Yeshwanth Reddy Yerraguntla: Yeah. Rajashekar G: like uh 78. Yeshwanth Reddy Yerraguntla: I Rajashekar G: So based on that we can fix like I am either way I'm Yeshwanth Reddy Yerraguntla: know Rajashekar G: okay. Yeshwanth Reddy Yerraguntla: there Enterprise project review. better. Oh, the UI UI Satyasri Prabhakar Mantripragada: I went pretty last time. Rajashekar G: Yeah. Yeah. Every week Thursday 1 to 12 Yeshwanth Reddy Yerraguntla: part. Rajashekar G: Mir Yeshwanth Reddy Yerraguntla: there. Rajashekar G: 1 to 1:25 p.m. Manisha Gundapuneedi: How much? Okay. Rajashekar G: every Thursday every two Manisha Gundapuneedi: Okay. Okay. Rajashekar G: weeks. Yeshwanth Reddy Yerraguntla: See problem somehow I will think about it and I'll try to fix it uh next two weeks. Rajashekar G: M Manisha Gundapuneedi: 7:30. 00:35:26 Rajashekar G: not okay. Yeshwanth Reddy Yerraguntla: Are they all first thing? Satyasri Prabhakar Mantripragada: Last half an hour connect. Manisha Gundapuneedi: Oh Rajashekar G: Is Yeshwanth Reddy Yerraguntla: Okay. Rajashekar G: it Yeshwanth Reddy Yerraguntla: 6 to 10:30. Manisha Gundapuneedi: goodness. Foreign Yeshwanth Reddy Yerraguntla: You can go on our Rajashekar G: Ganesh Yeshwanth Reddy Yerraguntla: channel. Rajashekar G: power. Manisha Gundapuneedi: speech. Foreign speech. Foreign speech. Yeshwanth Reddy Yerraguntla: Yeah, Rajashekar G: and Gopalaru architecture. Manisha Gundapuneedi: Um, Yeshwanth Reddy Yerraguntla: two teams take lead. So this will continue for does Manisha Gundapuneedi: foreign Yeshwanth Reddy Yerraguntla: not Rajashekar G: Anyone Yeshwanth Reddy Yerraguntla: repeat Rajashekar G: could Manisha Gundapuneedi: now. Okay. Yeshwanth Reddy Yerraguntla: and so on. One, two, three. Two and a half weeks. One week. Is this accurate? Rajashekar G: Just enough Yeshwanth Reddy Yerraguntla: I'm in a different time zone whether I put it wrong. Custom. Rajashekar G: Friday question. Yeshwanth Reddy Yerraguntla: Yeah. Saturday. Rajashekar G: Okay. Okay. Yeshwanth Reddy Yerraguntla: Basically, I have to put it from Sunday to Thursday for myself. 00:38:29 Yeshwanth Reddy Yerraguntla: Set. Yeah. Okay. Rajashekar G: Um, Yeshwanth Reddy Yerraguntla: Okay. Shall we have this discussion or what do you suggest? Rajashekar G: we can have it now. Yeshwanth Reddy Yerraguntla: Today we'll only discuss um uh what are we discussing? Rajashekar G: Corent line. Yeshwanth Reddy Yerraguntla: We have to commit to we need to understand how long the planning is going to take. Which means we have to understand what are the different things in each uh uh agent. You have to discuss uh who will work on which uh agent that is also I think important. What are apexics and we have general discussion? Maybe We'll try to discuss the simplest one available complex. So I prefer we take up either uh unifier agent or something else. Finally, we have to get a feel of how long will planning take commit to car. Rajashekar G: Oh, no. It's just one. Yeshwanth Reddy Yerraguntla: Sorry 9 10:30 Rajashekar G: Yeah, Yeshwanth Reddy Yerraguntla: 10:40. 00:41:07 Rajashekar G: sure. Yeshwanth Reddy Yerraguntla: So praas let's see how Satyasri Prabhakar Mantripragada: H. Yeshwanth Reddy Yerraguntla: we do this. By the way, Satyasri Prabhakar Mantripragada: Sure. Rajashekar G: Sigma make ID. Yeshwanth Reddy Yerraguntla: what the f***? Rajashekar G: Now we need to ask them. Yeshwanth Reddy Yerraguntla: Oh no. Satyasri Prabhakar Mantripragada: So Rakkesh Ping just he'll have that Yeshwanth Reddy Yerraguntla: Okay. Satyasri Prabhakar Mantripragada: control. Yeshwanth Reddy Yerraguntla: Yeah. We can drop off then Rajashekar G: Okay. Satyasri Prabhakar Mantripragada: So our target is till April 10th 6 weeks Rajashekar G: Thanks. Yeshwanth Reddy Yerraguntla: like uh Oh no. Satyasri Prabhakar Mantripragada: duration. Yeshwanth Reddy Yerraguntla: Oh no. Oh no. Midappril s. Satyasri Prabhakar Mantripragada: H. Yeshwanth Reddy Yerraguntla: Uh yeah. Rajashekar G: Uh I want to give one more update. So that's an equivalent environment where uh the Salesforce intelligency is deployed. Satyasri Prabhakar Mantripragada: Huh? Rajashekar G: Enterprise.gomi.com and chanika.gomi.com both are at equal stages right now. Satyasri Prabhakar Mantripragada: I don't know. Rajashekar G: So we have deventer enterprise.com Salesforce guit that is the current stage and mode environments. 00:43:08 Satyasri Prabhakar Mantripragada: understood. Rajashekar G: Yes. Satyasri Prabhakar Mantripragada: So let's continue using uh chaneka for internal uh testing dev will like that is a staging environment for Rajashekar G: Okay. Satyasri Prabhakar Mantripragada: us. Chaneka is our staging environment. Rajashekar G: Okay. Satyasri Prabhakar Mantripragada: Enterprisebrain.dewami.com is the production environment. Rajashekar G: Okay. Got it. Satyasri Prabhakar Mantripragada: Dev is your dev and asmin whatever is done on staging so far which is should be available on enterprise brain going forward. For now it is the other way around. Enterprise brain going forward all your Rajashekar G: You're right. Satyasri Prabhakar Mantripragada: dev pushes will go to chaneka and on chaneka once these are tested certified then they'll go to enterprise brain. Rajashekar G: Got it. Satyasri Prabhakar Mantripragada: So sales team will always refer to enterprise Rajashekar G: Got it. Satyasri Prabhakar Mantripragada: brain Rajashekar G: Right. Satyasri Prabhakar Mantripragada: URL doesn't make any difference but functionality wise we should Rajashekar G: Right. Satyasri Prabhakar Mantripragada: have uh uh everything ready on Rajashekar G: Okay. Satyasri Prabhakar Mantripragada: that Rajashekar G: And one one thing low we have found one bug gun actually authentication bug. Yeshwanth Reddy Yerraguntla: Mhm. Rajashekar G: So when the user is user is logging in first time Yeshwanth Reddy Yerraguntla: You think Rajashekar G: Gmail it's taking the tokens and all it's not asking for any login again it's directly jumping into the main page. Yeshwanth Reddy Yerraguntla: log out refresh. Rajashekar G: fixation. Yeshwanth Reddy Yerraguntla: You're saying uh it is uh Oh no. Okay. Rajashekar G: I'll assign it to someone today and get it done. Yeshwanth Reddy Yerraguntla: Okay, Rajashekar G: And we are PI is a dependency. Yeshwanth Reddy Yerraguntla: Marisha. Sorry. Rajashekar G: It's directly linked to pyog repository. Yeshwanth Reddy Yerraguntla: Yeah. Nice. Rajashekar G: Okay. Manisha Gundapuneedi: Oracle. Rajashekar G: Okay. Manisha Gundapuneedi: Okay. Rajashekar G: Good. Manisha Gundapuneedi: Okay. Mangala button. Bye. Transcription ended after 00:46:05 This editable transcript was computer generated and might contain errors. People can also change the text after it was created.