Feb 18, 2026
Enterprise Brain: AI-first Design Approach - Transcript
00:00:00
 
Satyasri Prabhakar Mantripragada: on a particular graph and how the initial con what information need to be shown to the user that is also not in our control but for example if I broadlyize them into if I want to show some summary section some data points to the user we are just telling that what would be the styling you assume that you are giving only the design system to them if you going you are planning to show a bar graph then use this thing if you're planning to show an sheet graph you show you're going to go with But if they're going with a plug-in, they're using some third party plugin and they because graph rendering everything is represented by a plug-in, then we have to give only the the color codes to it. What is a design system that we can recommend we can propose that that is the one aspect of it. The second one right the behavior of the page navigation because every for system interaction that user can make on the platform is an activity that user is performing. Sometimes the activity will update the same with a new or fresh data.
 
 
00:01:05
 
Satyasri Prabhakar Mantripragada: Sometimes the based on the activity the user is interacting it will take me to a different context all together. So the transitions how the screen because either it can be changing in the same canvas or it can be changing leading me to navigate to a new canvas. We need to maintain how we going to show the transition and view the features till you context. So because here we talking about only two things the navigation patterns we don't want to go with the traditional way of navigating the things we don't want to go that we need to find a pattern. What is another way to crack that and the navigation that's the one thing we know second one assume that there is an uh graph is generated when I and you are showing some KPIs or some graph values or plotted bars here. Now when I hover on this platform on this graphs right now we will only increase the um size of this bar or give some effect. So now assume that it it will also give a contextual query to me because I want to understand for example this is India this is Australia this is something and is is talking about one story.
 
 
00:02:30
 
Satyasri Prabhakar Mantripragada: Now if I want if I click on this my intent is that I want to go to the India. I want to drill down on India and we are telling that you are in the drill down version and for the drill down version you are seeing a different graph or different information. So assume that when I click on it with a slight animation, it is coming as a small thumbnail and it's giving a different view to you the graph. Now again it has a multiple entry points. Now I have a different KPA is talking about now I want to focus on this KPI. When I click on it, how we are going to represent this hierarchy of context that is going to the user the transition How we are going to advance this and paint this view to the user? What is the pattern that we can show to the user to engage analyzing generating and all that we need to show I told you we have to show some animation whenever it is Mhm.
 
 
00:04:02
 
Satyasri Prabhakar Mantripragada: I need a visual representation of something is happening. I gave example of Figma star. I give zoom in zoom out. So we have to give it next but it's a KPI but I want to know more about that history. So they assume that there is an interaction here where I'm contextually calling my a intelligent when I click on it how it is going to bubble it up to give that interface is giving an interface to let me know what you want and keeping the context of what I have selected and based on this now once the response is ready I'm not expecting with the content has to wake up. has to give a prominence to that feel of see one idea came to my mind. So whatever you are saying till now we did it like either with the back option or overlay. So how I am thinking is right What I mean is like I I understood. So how I am thinking we do the canvas.
 
 
00:06:11
 
Satyasri Prabhakar Mantripragada: Yes. I'm coming here. You can directly go to the response. If you're interested to have some mechanism just navigation thinking in a different perspective we are seeing what maybe you can give some thumbs up to this thing so that you know it will be highlighted in different way so that you know that I I like the conversation that what happened or this view I want as an uh note make as a separate view so that I can able to instead of making all these three drill downs I can directly make this as a query on my dashboard. Yeah. Got it. something like that to make it more intelligent layer and design beyond that I'm not able to think because I can able to drive the things from an a perspective what question need to be asked and what story need to be tell so we'll give one global level if we have multiple drill downs. Okay. Top low we have one thumbnail kind of.
 
 
00:07:53
 
Satyasri Prabhakar Mantripragada: Okay. If you click on it, it will expand. Assume this this is like a canvas. It is kind of over. So connection. We can give those kind of visual connections already. Whatever you said he showed us a demo what he has done is like you started your conversation somewhere. hardest regular. Even though different. So what he did in the show to me is like have a different thought I'm changing my opinion there is another branch is going to come So you started your conversations. I felt like that is also confusing to the user. It's like too many elements. research it's needed the results are going to change who is doing that hypothesis and all that that makes sense. will it be possible to like it not even be chart interface can automatically group this into okay financial results system can always take that learning over a period of time not on the first day.
 
 
00:11:25
 
Satyasri Prabhakar Mantripragada: first day, first week that you are interested into the quality because you are coming from an domain. You are interested in the quality inspection and all what is happening and all that. But later we realized that you are not in the uh that mode. You are more in the delivery manager. You want to see the success rate of the product. You always asking the questions related to why it was delayed and how can I make the make it deliver fast and to get a next number of profit. So we will change your perspective we won't show that the quality we will talk about the business impact only. So, so to give you a context to get an idea So it will be regular small. So I my thing is I am also struggling one side we are thinking that a for solution and we are limiting our things things to the aation the patterns the patterns when it comes to the pattern it's more about Now Uh okay one side the AI team is telling that you people just tell me that what we will analyze the user we will uh show him what is right information for him based on his question we know what to information how to present other side of the table the engineering team is telling that we will integrate with some charting tool or so already graph
 
 
00:13:45
 
Satyasri Prabhakar Mantripragada: So we are telling that what is the brand what we are primary focusing on discing ing that to break the regular truly interactions and all that just because people are used to having bigger scrolls and all that they feel like that is a way to consume the content and right now there is a relation between every content that you are looking into. Now if you are giving a focus area to the user and let him think and let him let him analyze what we are presenting and that is giving a new question to him. For example, in all the AF application, it starts with an empty canvas sending because always the the user has to put his brain what to ask the system. There is no contextual guidance and he also lost sometimes even the visionaries don't know what to interact the beginning. So now what we are telling our intelligent layer will not leave you at that position. So when you log in itself we know about you and we know what information that you prefer.
 
 
00:15:01
 
Satyasri Prabhakar Mantripragada: So we'll show the context of information to begin with and we'll given the potential follow-up questions that you might ask when you look into that data. So that is one of the trigger points to ask on that. We we are just giving helping him to take that but we were not sure whether our content and our follow-up questions will are helping that person also at the beginning because we we are we are assuming that will help and we are pro providing that now but before either he can go with the what we recommended content and he can further ask the questions on top of it to get that or he completely ignores that content and he can start asking his own question that how we will know based on the learning pattern system uh is uh doing in the background. We know that even though you you're telling that the person is using that lessly clicked only few times and most of the time he is diving from the the regular interface to ask the ask my system or these are the common question even he's going with that interaction thing these are the common question that he always asks so then our intelligent system has to take some time and tell that flip that instead of showing what you're showing earlier it has to do that so that that's the intent so for that what we have The
 
 
00:16:18
 
Satyasri Prabhakar Mantripragada: meeting might mute.
Rakkesh Yenugudhati: So, so far what we did and it's like we try to understand the archetypes archetypes of the person who are coming onto the platform. The archetypes are like a very high level a broad level category of the users because we are talking about uh our pyramid. If you remember we classified the entire user group into three segments visionaries and um executors not execution visionaries at the top level. The second level is the managers. The third level is the individual contributors. So we put the entire for a lack of better we are calling those things and we put them into three buckets. Now inside the three bucket we are there will be multiple personalities will be there. The people who are sitting on the visionary table how they will think and act how think and act how nav think and act is entirely different. Their perspective their aggression and all that are going to be different. So assume that those are the archetypes. Considering all the things we broadly classified them into 10 different buckets of the role.
 
 
00:17:53
 
Rakkesh Yenugudhati: We're calling them as a rapid decider, deliberate analyst and all that just to give a name to that but if you read the definition of it you will understand the Okay. So now if you see that I will go through one definition only so that you will understand sure. So the rapid decider is the first archetype we define. So we are telling that the person that who having that archetype is like they're fast, they're direction oriented and decision makers in the system and the core pattern is like speed and momentum. They always look for a speed and momentum in what they're doing and they they always prefer the conclusion first and the latency is very low and there's a moderate risk to all right that that is the decision style and what they prefer the trust by performance and evidence by demand because they won't believe because performance itself is the trusting the things and how they consume the information that what you're showing they prefer the executive summaries and then they go to the two drill downs and uh the that's the reason we telling that prescriptive is the first if you remember our analytical structure there is a descriptive layer there is an diagnostic layer there is an insight there is an predictions there is a descriptive but these kind of people always prefer the prescriptive things tell me the last word first what says right that's what they prefer and they will
 
 
00:19:37
 
Rakkesh Yenugudhati: slowly they will drill down to the the need and they highly they prefer the automation of the things and where uh they starts with the perspective layer and show the impact and confidence and expanded the evidence and one click execution. This is how the persona archetype is defined like that. Every persona has its own for example you take the deliberate analyst they want a because it's deliberate you want a structured a logicaldriven approach and uh they they validate before they commit and they explore first and risk averse. So they explore they they want to try first they want to see the all the possible opportunities they need a proof. So what they consume they are the diagnostic tables they need a tables to understand the complete risk they're involved in it and their competitions and then only they will take the actions on it. Uh if you see that the enterprise how it is like they show the drivers they will show the drivers to them they will take the assumptions and then they follow the recommendations and display the methodologies. So this is how they will because we are just bringing the the high level personality here.
 
 
00:20:42
 
Rakkesh Yenugudhati: Now even the rapid deciders because even Navin is also rapid deciders but the the the speed at which PMA takes a decision the speed at which Navin takes the distance are different but both are the rapid deciders only come to their own context and all that. For that what we have done is like we further categorize this into personas. Each archetype was was break down into multiple persona. Even if the rapid decide is an archetype there we have defined the personal like the velocity driver the pattern commander the constraint cutter but don't go with the names but see the definitions of it. What we actually mean is like the makes the fast calls to keep the org moving that tolerates the imperfection info if the direction is clear and whereas a constraint cutter like remove the blockers aggressively simplify the complexity into fewer layers and what are the core traits if you see that their conclusion first they are the simplifiers they want to make the conclusion they are the antifures so like that every these core traits will define the personas but all these are falling under the rapid decider archetype only.
 
 
00:21:51
 
Rakkesh Yenugudhati: Now what we have done is like we tried to map these personas. I'm just explaining one line with you the context here. What we have done is now we know the archetypes. Now we know the personas. Now we are taking that our benchmarking three things the the visionaries and executives the manager and operational individual contributors for those things which are like dominant. So if you see that the rapid deciders who are in that role the visionaries are more dominant there in in few archetypes are other personas are dominant. So like that we are able to differentiate where how it is going to help is like for example now when the user is onboarded ID before I go to the I also explain how we are going to categorize a person who logged onto the system into particular archetype bucket is like considering these attributes by looking into because at this beginning of the things in this entire attribute list you might be having only few and because you are be integrating with the system. Whatever is possible that with the basic information that what you have you are identifying the person is potentially will be in that in that bucket.
 
 
00:23:06
 
Rakkesh Yenugudhati: I repeat that. So there also now these are all if you there are 20 30 attributes we defined and but we tell you that all the attributes will come when this person is start interacting with the system and when he performing activities on system we know more about that person this will be increased. Now on the user first time login assume that we are integrating with only Salesforce and uh other so now because assume that this is like their uh decision making capabilities I'm just drawing the one action is like decision making capabilities the other action is like time in which he is going to take the decisions I'm just putting as but Assume that there is an uh different archetype buckets are there lies who are like take time thing and then all the 10 archive buckets will be displayed here. Now in the same archetype we have a personas right we defined that four to five personas. Now initially I thought a person will fit here in this archetype is what we identified and that also a person belongs to this thing.
 
 
00:24:26
 
Rakkesh Yenugudhati: Now I know the traits of the archetype. Now I know the core traits of the persona based on the system is recommending something when I log into the page. I don't know what we are recommending under that system considering that system is going to recommend something to me. Now the user based on behavior he's interacting with this or his inter based on how interacting with the agent and how detailed he that he's going on the content and all that the attributes are getting matured for the persona then the two possibilities are there what are the two possibilities is like one the mobility of the persona moving that initially I put this in the archetype the chances is like the persona the archetype is right but the mobility is like he has to not this persona he is this person so the mobility within the archetype itself is one mobility the other one is like I initially thought this thing but he is a this guy this is the archetype so system that maturity level changes based on that you know what kind of visualizations they prefer you know how deeper they go into the content that's why we can to hyper personal systems so this is one exercise we did to get a clarity and depth of what we are doing to make it as a scientific approach we not randomly jumping on it.
 
 
00:25:41
 
Rakkesh Yenugudhati: First we define the archetype and then we map the archetype to the personas and how we are going to define that each archetype and persona has it core traits to identify the user and then we are going to put them in the initial landing page based on that. Now the second exercise that what we have done is like we we categorize similarly how we did the archetype for we considering the users on the platform we take the archetype for considering information as a thing the information archetype is what we consider so information archetype more broadly we brought into six categories of it the one is like event specific time series I will show you evententric what that mean by evententric is like is always about a thing or an attribute. You're talking about something and it's always a tuple. It has it own thing. It's row and equalent thing and it's a relatively stable because it's a tle. There is a change in fluidity is very less on that and visual patterns typically we recommend for those things is like table bar charts and all that and why this work.
 
 
00:26:44
 
Rakkesh Yenugudhati: So we are trying to understand these things evententric the state statecentric time series distributional relational spatial text thematic these are the archetypes in it and what is going to differentiate that because it's a thing and attributes and occurrence and action it's a condition and snapshot and what is a potential way to represent that. So we try to understand the archetypes here. There we understand the archetypes of users and we map it to the personas there. Here what we we understand the archetype of the data information and then we we try to map it to the data. What I mean by data is like data we have two types of data the qualitative and quantitative. Now in the qualitative itself we quantitative itself we have some the multiple data things. It can be a discrete value. It can be an internal iterative uh ratios, percentage and all that. These are all the quantitative and similarly we have a qualitative which talks about distribution data, probabilistic data and all that. So here we try to map that to the archetype because if you're talking about this this is an entitycentric.
 
 
00:27:48
 
Rakkesh Yenugudhati: So now we have a clear clarity on that if you are able to if the user based on the question if your data is falling under the entitycentric uh archetype then then potential the data will belongs to any of these things or the whatever as the data first assume the data is coming from your out your query and system response if it is falling under any of the category then it belongs to entitycentric if it belongs to entitycentric then what we have what is the right visualizations to show them. So if it is an interval data we are telling supported visations are line chart but if you're thinking about an AI thing then what is the contextual comparison table that you can give to the user and uh and just we we try to map those things and so now we have a clarity on the users and we also have a clarity on the information and the data. If this is coming as an information and this falling under the particular data type and particular uh archetype then the potential visualization that you can bring onto the table are these things that is the other mapping that we have done.
 
 
00:28:57
 
Rakkesh Yenugudhati: So in this case in the previous sheet when you mapped data to specific types of representations do you also rank uh these representations that also because learning is always there because we here the data is like a very scientific approach and no matter what the KPI will be a KPI it's a single number and the record is a record so here the the learning will be there but I'm seeing the learning will be slightly I mean the changes will be the mobility will be very less but in other case we are guessing the persona because once the system is matured there so there the mobility will be very high so the user is able to see the changes here is more about the same information can be represented in a multiple visualizations but he is based on that but for example Andy if I read all the conversations that what chart gives to me at the end what I ask is like you make comparent table for me but quickly analyze and compare So that is my pattern for anything. If the system knows me that pattern then instead of showing another way of representing it always stop me at table level.
 
 
00:30:03
 
Rakkesh Yenugudhati: So instead of saving one more query that I'm going to ask on the system. So like that we their system is going to understand not from my personality it is a potentially I can able to represent the same data in three different visualizations but as a Rakkesh prefers this. So for me when I ask it will give that visualization. when you ask it will give a different visualization because it's a potential possibility. So so far we arrived at this point and because it's two ways to understand because now we are talking about where to put how to put and what to show the data we felt that if this information is available and if we want to talk to them how we arrived at this point the HDL layer we are talking right that is this is nothing but the HDL layer we trying to build for the enterprise brain now with this data I know that Rakkesh is not is not that aggressive is like he want an evidence before he take the call and he's heading the the particular department there and responsible for his activities.
 
 
00:31:02
 
Rakkesh Yenugudhati: Now when he come onto the system what kind of information that he's asking he expected a day briefing to him he prefers a debriefing but with the vener coming onto the platform and he is like a very aggressive in taking the calls he don't want the debriefing he want to have an a tell me the bottlenecks that I have to fix today he's expecting that so if you show me that information up front by this scientific reasoning then our system will have a hey is what the thing so far I this data hand now we need to think about the whatever I discussed right now in terms of to bring that visual way of doing that it's like is more about the two things now the a engine is going to take the call where to put the person what visualization what data need to be pulled that is the a engine is taking but how to represent that is taking the front end engineering part because they're going to integrate with the D3JS or D3 charts or high charts or flourish uh they're going to take that.
 
 
00:32:03
 
Rakkesh Yenugudhati: So now in that context our where the design can fit well is like telling that what is the branding that we can give onto the what they're building. The second one is like what is the navigation and interaction patterns that we can able to bring onto the table to give that it's an AF first product when I talked about the navigation right every so far all the applications that what we have seen as AI it's leaving at a canvas we don't want to leave at the canvas the user when you log in that's the one thing and how we are going to show the user that contextual information and we don't want to load the too much information also just one piece of information that we believe that he will work and potential things that he can also have. It's like an uh quick options for him to trigger the question directly. Now the two ways the user can able to interact with the system. one he he's convinced with my summary or debriefing and I I want to perform a certain action on that either I can interact so that it can drill down system can build down to the next level but when I build on to the next level in a traditional system they introduce a breadcrumbs so in our A for system we want to go away from the breadcrumbs but the context but the same experience we need to bring in the breadcrumb is helpful to maintain the context from where I navigate here also we need to maintain some
 
 
00:33:25
 
Rakkesh Yenugudhati: context here that is the one the second one right uh is more about how you are opening the uh the conversations for me the because I'm convinced but there is no drill down further but still I have a question on that how you arrived that number is what I want to understand contextually bringing the AI agent onto the top and asking that question how that AI agent is coming coming onto the platform and how he's interacting is the other thing that we have to do so it's more about how the experience of the navigation how the experience of the user feedback that we can give to the user when he's doing the certain things is what we are so we have to crack those things but we have some ideas but we need to spend some time to do that so one thing one current our current priority is Salesforce integration where it's all about numbers deals and uh all the progression the other scenario is where enterprise brain is integrated with other systems like Gmail will work nothing logic is not going to change it. So this one so what I wanted to ask is in case for example there's a deal that is stuck in a certain stage for more than some time for the same customer for whom we are already serving on an existing live project we have a project X now this deal is for project Y from the same customer but that was not moving beyond a certain stage for last one week or two weeks now the CXO wants to understand why is it stuck there so when for example
 
 
00:34:59
 
Rakkesh Yenugudhati: this is my assumption that when I hover on it or I wanted to get more information would it immediately get information from my say Jira or my uh say emails and so on other systems where it can highlight that okay there is this risk the previous project from this question that's the AI problem for me so I am asking from a visualization standpoint yes because for me right for example Assume that it's possible I'm telling what we are going to tell whether you want to convey that message in a form of a a suggestion or potential action or as an information only these three ways right if it is an information whether information can be a qualitative information here the quantitative information won't come because here telling that you because already this customer is with us from working with the same customer from the last seven months onwards is an actual customer is a qualitative data M so we we already know that how to represent the qualitative data. So the system will take the intelligence of showing that for example you are telling that the python this is not moving and you want to understand why it is not moving.
 
 
00:36:12
 
Rakkesh Yenugudhati: Again it's like it's supported by qualitative and quantitative data. It's a combination of both. So system knows that if the combination comes how to represent the data. So we we are telling that how to representation is them. So I'm not worried about what is going to come previously right we are worried about what question that you are asking how many drill downs it will have will be two or three and all that now because it's the system is taking a deterministic pages we previously it's nondeterministic we used to turn deterministic pages right now they're telling that no we won't be able to make that deterministic it's a nondeterministic thing so it's dynamically like if I want to make a five drill downs I can make it or I can limit to the three drill downs they can able to do that the system is taking that call so we just we need to support what our system is doing. The system is not breaking anywhere. Okay. So for and uh so okay that that answers to some extent.
 
 
00:37:10
 
Rakkesh Yenugudhati: The second thing that I was thinking is if you are giving a drill down from one graph one image some kind of a graphic um to another graphic and so on finally would you land on it simple text to give the information it can be a simple text or it can be a visualation with the text. Mhm. So for example and it depends upon what question you are asking because that question is always I can able to put it in a simple quantitative sentence I can able to do but in order to support that quantitative sentence I need to show some proof to you that can be visualization along with the quantitative information that I'm okay yeah I was coming from a standpoint of like you wanted to present don't want to say this but kind of a breadcrumbs a linking from a parent context to the current context you wanted to have some kind of a hook or some kind of a link. So if it comes to be like that with a ending to be a text based answer at the final how would we trace it back to the initial hook because that's the reason that either you whatever you show that some will be there we we are telling that how you arrived at at this point of time you in order to get that whatever the final summary you have went through this thing but potentially without going to that route also you can able to achieve that.
 
 
00:38:31
 
Rakkesh Yenugudhati: So every query that you asked can be an independent entity in your you can you can mark that as in I need this on a daily basis and you can like how you are putting your widgets on your phone right what you're using you can put that widgets directly onto the your dashboard because initially we are recommending something but user can preferences so the system automatically will pull that because it's not like a it's not an API right having constants sing there literally user has to hit on it to get the data. Yeah, because only that is only the difference and in the previous uh applications right it's all APIdriven whether it's a continuously either a push or a pull kind of an information here it's always a push only and pull pull push because how the API is defined. Yeah. Nothing is going to be dynamic there. So we need a little strategy or work on this particular I need to have some will be more maybe I know that but Just next we were not even there because now I'm able to talk to this thing because I because of this work what we did but this will be the input for the AI to make the right decisions.
 
 
00:41:20
 
Rakkesh Yenugudhati: But I am able to land him based on some conclusions and I will improve that person to a different archetype and persona. Yeah. I'll choose some and we will uh we not telling that I know the importance of it but literally we're struggling on few resources. Okay. He said that he's not well. But anyways, we are recording it so far. If you guys want to explore this, I can share this file. It's too much for you. I will keep it. Okay. She may call it as shortterm but the this is what our final product is going to consume but for immediate she needs some presentation layer and maybe once we I can able to understand how can you how can I prioritize the things even know that we need to work on what they're doing but interactive graphs we already the flourish thing particular two aspects that I look at is one interaction on a graph I can click on any random point and ask for information you assume that every point that you're seeing on the screen has an API definition there because it's a it's just you're passing that parameter dynamically to the back end is going to give some response to you to show to the user.
 
 
00:43:44
 
Rakkesh Yenugudhati: So assume that every point that has on the screen is clickable. We have that infrastructure if not ready can we build because the capabilities that showed me everything is possible kind of a thing for me. He touched every graph node and told me that I can able to click. I can able to change the certain parameters and all that. That's what I was thinking. JavaScript. Javascript applications. I know that why it has an ambiguity before I want to put some friction to the user because based on that we can able to do Yes. So transition context putting at certain position in certain way. M so which looks like a very small small but it will enhance the platform we have to literally show some we can't go from there to make to understand what we are talking. Okay. decent level and it's not direct. UX minimum I mean yeah Only concept noneterminist Net behavior modes of layout we will start from there.
 
 
00:48:58
 
Rakkesh Yenugudhati: For example, increment Independent templates. For example, kind of charts. two types of interactions down on the same you have more context to ask for example India the revenue of India GDP of India I can ask about any other questions specific to India. Tell me the revenue growth of India or tax percentide. within the contentually But at least to get started, right? That is an intelligent layer. Another intelligent layer that we have to build actually. very straightforward because for me it's like one more query you are passing to the system to get a different response for example I started with that show me the leaderboard of GDPA for Asian countries for example Same question. two irrelevant activities by default countries. So Okay. Show me the tax percentage of JPEG that there is a a symbol for me to invoke my assistant. It knows that context of GDP concept then I will take the user to that layer.
 
 
00:53:43
 
Rakkesh Yenugudhati: some potential actions that you can ask. You want to view the state I will show that how to show that symbol on that particular graph. First question patterns for me. GDP third region wise GDP region wise GDP Only which can support that first single countries to give that So that at least we have some points. But for example, Right. That is one single I have multiple options to show like you know different departments different departments. Okay. Out of these 100 people 20 belongs to 20 belongs to 20 belongs to Right. Uh represent column clusters clusters is a challenge for me. will enhance. I will show another thing. I will show you. It's asking question and health and momentum. So to use that scenario you are expecting two different colors to represent that finerfed is coming from a very at least I'm not for Example one already.
 
 
01:00:40
 
Rakkesh Yenugudhati: Hello. Where are you sir? Yeah. Rakkesh party meeting. Sorry. Byebye. What the f***? Bye. See you. Bye. Only 30. 30 to 12. You have to pick the five shades of that color to categorize. There should be someone who understands then we already proven drop down is a drop down at the end. It's already Because of example, time series compatible graphs. Alternative So we can tell that how do I transform from this experience to that experience experience. experience. Bubbles open experience. Two different queries. I know that if I asking if the potential visionary guy is asking this question the chances of asking this Revenue play something sales revenue based on the department. Next department are already back. I assume that first question to respect. For example, second I feel like this is what I can able to do. already applicated as intelligence to system and that will dynamically pick it putting on to his bucket. But only the person has to confirm because has to take the responsibilities in the way that has to tell that this is good enough for me to do.
 
 
Transcription ended after 01:09:43

This editable transcript was computer generated and might contain errors. People can also change the text after it was created.