Role of AI Agents in Digital Transformation

Yash From Momentum91 (00:00)

Okay, it says that we are live. Now we don't know whether we are and if we are then for all the people who are watching this is the first time that we are using a new platform and so we just want to make sure that we are really live and so Koushik is checking whether we got a feed going on onto a profile.

Koushik From Momentum91 (00:07)

Yeah.

Yash From Momentum91 (00:27)

Then once we get a confirmation that the feed is ongoing, we will start the session. Not yet. No, not yet. Not yet. OK. It shows 40 seconds. Can you check on YouTube?

Koushik From Momentum91 (00:35)

good.

Yes.

Yash From Momentum91 (00:41)

Yeah, just, okay. It says that it will start soon.

Jay From Momentum91 (00:47)

I guess we are.

Yash From Momentum91 (00:49)

Perfect. So we are live on YouTube that I know for fact. I'm not sure about LinkedIn. We'll figure it out now. But since we are live, we've got to take this on at least one channel. We've got to take this forward. So we'll start the session. So hello and welcome to Momentum Officers. My name is Yash and I'm joined by my co-founders Jay and Koushik to discuss topic of the week.

Yash From Momentum91 (01:14)

role of AI agents in digital transformation. Our goal with these conversations is to provide you with actionable insights and practical strategies that you can apply to your own business. Throughout the session, we encourage you to engage with us by asking questions and sharing your thoughts. This is a fantastic opportunity to learn from each other and gain new insights that can help drive your digital initiatives forward. So let's get started. Jay, Koushik, how are you doing today?

Jay From Momentum91 (01:40)

doing it.

Koushik From Momentum91 (01:40)

Good, good.

And we are live in LinkedIn,

Yash From Momentum91 (01:43)

We're live on LinkedIn also? Okay, that's good to know. so finally, so the platform works. We know that there is some very simple basic functionality that is working as expected. But I know for a fact that Jay is excited because he's going out for a travel and Koushik is relieved because he's just come back from one. And so Koushik, can you talk about how was your trip and then we'll move to Jay.

Jay From Momentum91 (02:06)

Yeah.

Koushik From Momentum91 (02:12)

Yeah, it was great. So we had a meeting with a client. So usually all the clients that we work with, we have this phase where we try to understand their business. try to see how the implementations or the platforms that we build for them could go about. So in depth, so we sit with the entire team. So we had like...

Like every single, the entire week we had meetings and every day we had like five to six hours of meetings. We calling department by department and we sitting with them and opening all their books. And you know, seeing their legacy systems and trying to understand how they have been currently doing. So very long meetings, I think, but yeah, it is all it is.

Yash From Momentum91 (02:46)

Yeah

Yeah,

it's an interesting digital transformation project that currently Koushik is working on. And Jay, you are going on a different kind of a trip and that we can say from the smile that we see on your face. Talk about that a little. Why are you going? What are you doing?

Jay From Momentum91 (03:11)

yeah sure, just thought to have some adventures during this year and to get started with planning to go on a trip where I'm learning to surf.

So I'm going to one of the well-known places in India for especially for surfing it's also will be staying there in South school only and Morning, it's like from 7 a.m. To 11 a.m. You are learning how to surf and remaining part of the day you are working Maybe in the evening if you want to explore the city you are doing that as well So that's where I'm coming from the smile that you see today is yes because of I saw sunrise while staying doing the packing because was working last night, but

Yash From Momentum91 (03:40)

Mayfair.

Hahaha

Yeah.

Koushik From Momentum91 (03:48)

Nah.

Jay From Momentum91 (03:48)

full energized

for work plus learning new skills.

Yash From Momentum91 (03:52)

for the,

for what they call, workation is what it's, it was popularly known as. But interesting, yeah, surfing is something that a lot of, I don't know, like you are the only person that I know is pursuing surfing. Yeah, interesting. So coming to the conversation that we want to have, right, which is essentially the role of

Jay From Momentum91 (03:55)

of holy.

Koushik From Momentum91 (04:03)

It was us. Exactly.

Jay From Momentum91 (04:05)

Well, am trying it for the first time. So let's see how it turns out.

Yash From Momentum91 (04:17)

AI in digital transformation and largely will be focusing on AI agents and what they are doing for organizations in different sectors, primarily retail, distribution, manufacturing, because these three sectors are something that we understand really, really well, having deployed a few projects, worked with a few clients, understanding a few use cases. But before we go deep into the conversation, Jay, if you can explain to us what is an AI agent,

I mean, what is like, I, we came across like machine learning and natural language processing, and then we had AI and then generative AI. within that, we were talking about large language models and all of those things. And then now we've been hearing about AI agents as well. So can you give sort of a brief around what are AI agents and what's the, what's the general use case for those?

Jay From Momentum91 (04:54)

Great.

sure.

So let me try to explain it in a very simpler format. All of us are very much aware of generative AI where we are asking certain questions to chat GPT or any other model. And based on that, we getting some answers. Many people are using it for creating some form of content or drafting some emails and things like that. Now, think of it this way that this generative answers are there for sure. But along with that, if there are

certain set of tasks which are supposed to be done. for this, so basically AI agents, can say typically are an autonomous system, which uses machine learning, NLP and data understanding of data. And it basically does some predefined tasks.

The advantage it has or the difference, so like many times people do confuse AI agents with automation, right? If there is some automation happening, it doesn't necessarily mean that AI agent is working on that. The preliminary difference would be here. There is some amount of data-driven decision-making that is happening. So AI agents are preliminary autonomous systems which are doing some tasks and they are using machine learning NLP and...

obviously data understanding based on which it will do the task. As we go along, I'll share across some examples which will give the clear definition and differences between what is automation and what we can consider as AI agents and how they are significantly more effective in getting things done.

Yash From Momentum91 (06:35)

Correct. And Koushik, while you ask your question around AI agents, if you can look at the screen and make sure that your face is visible, because it seems like you're a little far away from the camera and part of your face goes behind your name. So if you can just adjust the frame, that would be great. Perfect.

Koushik From Momentum91 (06:51)

Yeah. Yeah. So I was trying to understand Jay, like for example, let's say if I'm a business owner and I'm, I have multiple departments in silos, right? So how, how do you map what use case according to my business and how would you go about, like, what would be the blueprint for this for me? If, I have to go about implementing this.

Jay From Momentum91 (07:11)

Right. So for a business owner who wants to, you know, utilize AI agents for the digital transformation, initially, the very first step would be, I would say, to understand what are the pain points in the current processes, right? So every just similar steps to what we have in I mean, digital transformation in general. Now in that, once you've identified those pain points, what you need to do is then start with

like low risk AI automation use cases, figure out what are those. And then based on that, you need to also ensure that AI is also aligning with the business goals. And then you need to, I mean, leverage some form of expertise. So let me go step by step into this and explain on what and how this needs to be done. So first you understood what are the pain points and what needs to be optimized. Based on this, you'll figure out that, OK, for this set of use case,

you need to understand what model will be most useful, what sort of architecture of the model will be useful for the same. And then based on that, you also need to have well-set other infrastructure. it's not that we inherit AI agents in the system and things get done.

There are lot of other things which needs to be figured out when we are talking about using AI agents for digital transformation. that would preliminary include a form of collection of data. So I would say customer data platforms need to be set well, where data is getting captured from multiple places. All those data once it is gathered, it should be passed and connected through API with AI. The AI model understands this data.

AI model also needs to be trained upon this. So largely, I would say first collection of data needs to happen from there. You need to integrate AI so that it understands the data and then you need to train AI so that it knows how to predict or how to do certain set of actions based on what use cases and along with that, need to also so if it is related to user engagement, then you also need to make sure that you know you are having systems ready for

getting user input along with that most likely you need to understand what sort of preference is need preference needs to be there and you also need to understand where like contextual personalization is happening. largely understanding of model understand having a knowledge of you know what model to use and how to train the same and how for it how should it interpret the data. So for people

who are business owners who might not know well with respect to what sort of AI model is there, which is best suitable for them and you know how to integrate things. They just need to, I would say one should consult the partners, but then these partners needs to be evaluated in form of what sort of understanding they have with respect to use case models along with that how they are, you know, managing the system which I mentioned. So yeah.

Yash From Momentum91 (10:08)

So let's take up a case, And so that we are able to sort of, think highlighting and understanding the difference between what automation can do and what AI agents can do is a little important. So let's take case of a client that we have. So we have a client who does office space interior design, like commercial real estate design for offices that are 10,000 square feet or higher.

Jay From Momentum91 (10:21)

Yeah.

Yash From Momentum91 (10:33)

And so the process that they have is once they close a deal, they will do a 3D design for the whole, let's say 10,000 square feet space. They'll get approval from the client for the whole space. then they will have to, so once they get the approval, they will have to start building the office. And one of the key things that they need to do is to prepare this thing called DOQ. So how much glass is going to be there, how many chairs, how many tables, how many air conditioners and fall ceilings and

raceways and all of those things. So all of us can relate to working in an office and all of us see the things that we use and we are sort of surrounded by these things. So I think this example could be helpful. In the process, what they previously used to do is a person would look at the 3D design and figure out, so a person would look at both 3D as well as 2D design and then figure out what is the bill of quantity.

how much amount of glass panels, how many doors, how many washrooms, how many tables, how many chairs and stuff like that. And then they'll prepare it. They'll send it out to different vendors for procurement. They'll look at the quotes and then they'll have their own method of selecting the vendor. It's not always the price, but they'll have their own method of selecting the vendor. And then they'll procure the materials and start building the office. In this particular example, what

is the thing that an AI agent will be able to do or achieve reasonable amount of efficiency in this procurement process for an office of 10,000 square feet that is being designed by a client.

Jay From Momentum91 (12:05)

Yeah, right. So to answer that, let's try to first categorize AI agents and then we come to this answer, right? So I would say there are certain types which we can define the agents. What could be?

a rule-based where it is just following predefined rules of JODAN. Second would be machine learning agents where it will be learning from data and then it will be improving over the time. The third would be conversational agents, which we basically consider AI chatbots, virtual assistants and likewise. And the fourth would be autonomous decision-making agents, are basically doing some form of AI, a lot of use cases, but AI predictive maintenance or fraud detection

or based on the industry, these could be there, right?

Now, for the use case that we are talking of our client, it's mostly related to automation where things are predefined, where we know that these are the set of materials and there will be pricing based on some vendors which we already have. this is a good example where it's more of all the database are well stored in the system and based on the use case, this thing needs to be generated.

But how AI can efficiently do this is more based on scalability. So uses of AI agents can be multiple. It could be to improve the scalability or to the efficiency or to handle the large data sets here. If the data sets is very large, if we see that if a human is basically doing this activity where you know.

As you mentioned 10,000 square feet or higher. So what happens is a lot amount of data needs to be figured out. What are the items? If you try to prepare things based out of it, a human when it's done, it will take a lot amount of time. And then sending this data. So you may use some automations which are just, you know, where a human intervention is there.

but still they have to pick and fetch certain data and then based on that they create a BoQ and then ask for the quotes and things like that. What happens is this process can be optimized by using AI agents where you know they

typically understand what client is needing based on the initial discussion and where the preliminary conversation happened. AI agents are going to fetch the data based on what discussion happened with clients, what sort of the requirements are they will be able to predict that okay, this is the requirement that it already has created user personas in backend. So it understands that okay, this set of clients, it has been a repeated pattern where they might need these set of materials or these set of items. So it also includes things like that in BoQ and then

it passes on to human. sorry passes on to vendors. So what happens basically the difference is clearly visible. One thing would be a scalability. If this if there are a lot of requirements and then there is a limitation of people who are getting it done also a lot amount of time gets used. AI agents can do these things significantly faster at a very larger scale because they can handle this data and more importantly there will be less amount of errors in terms of

Once the model is trained well, will be less amount of errors in terms of predictability on what sort of personalized things need to be given. So hyper personalization will be even more effective in terms of delivery. And that's how the overall organization can deliver great value to their customers.

Yash From Momentum91 (15:19)

So what were the, sorry before Kaushek jumps in, so what were the four categories that you said? You said conversational, you said decision making.

Jay From Momentum91 (15:29)

Yeah, so sorry. So the first one would be very rule based that is just following the predefined rules. Second would be machine learning agents, which are mainly towards just the learning part of it. So it will be learning the data and it will improve over the time. But then it needs to be connected with other agent to get things done. if we're talking about multi agents being connected with each other and getting things done. the third one was conversational. And the fourth one would be like autonomous decision making agent, which is basically

Yash From Momentum91 (15:33)

rule based.

Hmm.

Jay From Momentum91 (15:57)

you know, fetching the data from machine learning agents and then it will be taking some actions. It will be more related to decision making.

Yash From Momentum91 (16:05)

And so would this be a good framework to think about, right? Which is where, you know, like a business leader could look at their business and identify what are the pieces where people are having conversations that can be had without involving people. And so those pieces can become part of conversation. What are the pieces where, you know, very basic rule-based decision-making is being done by my people if I'm that business leader, then by my...

Jay From Momentum91 (16:12)

Yeah.

Yash From Momentum91 (16:33)

by my people. I would sort of, so that's an interesting way to categorize all the processes that I have in my organization and think that, these are the places where significant amount of time, effort, energy is wasted and it can be made more productive using AI. Great way to think about it. So, but Koushik, you had a question.

Jay From Momentum91 (16:46)

Yeah.

Before Koushik asked the question, as you mentioned that, so this is very interesting, right? Where the business owner does not necessarily need to think on where the human part can be just replaced. That is one way to look at it. But the more focus is not just on replacing, it will be more related to aiding humans for getting things done faster, right? So it will be like AI augmented work, but.

Koushik From Momentum91 (16:54)

Yeah.

Yash From Momentum91 (17:14)

Got it.

Jay From Momentum91 (17:16)

The preliminary ways to look into this is one, the scalability part. The second would be the cost efficiency. The third would be improving the speed and accuracy. Because in decision making, also, good amount of accuracy is needed in terms of understanding the data and getting used to it.

Yash From Momentum91 (17:32)

So same people

can deliver more value. Got it.

Jay From Momentum91 (17:35)

Exactly. Same people can deliver

Koushik From Momentum91 (17:35)

Yeah.

Jay From Momentum91 (17:36)

more value. Obviously it has its own challenges on the compatibility using technology and things. But yeah, I mean, it's more about aiding the people and not replacing.

Yash From Momentum91 (17:47)

Got it.

Koushik From Momentum91 (17:47)

I think I can add to that because I'll give a live example from the trip that I just came back from. So we were discussing about the entire digital transformation for this company and we were talking about how we could do about dealing with their finance and accounting module and doing things with that. And let's say we were discussing one aspect of cash reconsilations that happens across. Now they were telling me that they currently have a 10 people team.

who is sitting and just doing cash reconsolations and it's the same task that they are doing every month. Now they want certain sort of automation setup within the ERP systems that we are trying to set up for them. But even then, at certain business levels, will be ERPs could do only probably 80 to 90 % of it. There's still the 10 %

Yash From Momentum91 (18:18)

wow.

Yeah. Wow.

Koushik From Momentum91 (18:40)

of it, which still needs to be done by human because there are multiple factors involved. I could like we could say that, for example, let's say there's still a 10 % that is left, right? Like for example, there's a printer that they're buying and let's say there's an ink that they're changing for the printer that doesn't come under asset that comes under consumables. And like when I have like multiple consumables like that, how do I reconcile them?

that becomes a problem, petty cash becomes a problem. So all these things, which is very specific to retail industry. Now the problem is for the finance and accounting team, 90 % accuracy is not perfect. It has to be 100 % accuracy.

Yash From Momentum91 (19:15)

It's not

good enough, It's like healthcare, right? If I save 9 out of 10 people, it's not good enough. As a doctor, I've got to do more. I've got to save 10 out of 10. So for finance also, that's not good enough.

Koushik From Momentum91 (19:26)

Yeah, it's not good enough. 100 % it has to be. and they say, and they have been telling us that how can we solve this 10 % thing? Because there is no way even among the most advanced ERP systems to solve that 10 % because it's very specific to their business. Right. So that's when these AI based agents would come in place, right? I could create an entire prefixed code paths or workflows and orchestrate them.

Jay From Momentum91 (19:44)

Yeah.

Koushik From Momentum91 (19:56)

So that I know that this is a repeatable task that is happening only to this particular business. So, and it could talk back to your ERP system and you you could have it still, you know, logged and registered. So.

Yash From Momentum91 (20:07)

But if they,

Koushik, if they have nine people team just counting cash, are they underpaying us? Like that's my takeaway. If they have nine people counting this and this, but jokes apart, right? This is the, this is the state in the world's most finance forward economy, right? India has like the most finance forward economy where digital transact, like half of the world's digital transactions happen in India.

Koushik From Momentum91 (20:13)

You

everything right now.

you

Yash From Momentum91 (20:36)

half

of them, right? And so if in those places also you need to have separate teams, I'm sure even with developed economies, this is a bigger challenge.

Koushik From Momentum91 (20:52)

Yeah, but I had a question around this day. So for example, if I'm a larger, so currently the way as of BC is that if I have to have some sort of AI agents implemented for my business, should always, it's a matter of cost that comes in front of me, right? And if I'm a small business owner, do we have any tools or anything in framework currently as a small business owner that I could?

refer to because there's also lot of clutter around the topic there's no clarity at all right for me to understand what it is so that I can take informed decisions upon it because so how how so first is what resources as a business owner I can refer to or go to one so that I can get a clarity of thought which is not so technical but enough to make me you know make decisions about who should I reach out to or go for

Then second is that like what do are there any sort of platforms that exist that I could you know use or anything of that sort?

Jay From Momentum91 (21:52)

Right.

Quick answer from my side because we are talking about business owners which are into retail industry, manufacturing industry distribution or anything healthcare for instance as well. Now these people are really good in doing their business understanding it. I would not recommend them to look into any particular tool and see how things can be implemented in their own system. If they want to understand like how agents work, they can go for like taskkit.com or there a lot of things available in the internet to just understand what's how

What are the capabilities of AI agents? Taskkit is very basic example of you can just create your own AI agent, very small one and get things done. So for example, if you want to just write down any topic on LinkedIn regarding any particular thing, what it will do is it will actually look into internet, get things researched based on this. will just create a post and everything is happening in front of you. So you will be a little surprised on how things are done like this. This gives you an idea on the capabilities of AI, but then

And again, coming to the question, I would not recommend them to just look into any specific tools because see this answer is very industry specific. And also, rather than this, they might want to look into any partner who is specific specifically into these systems who are having a good understanding of AI models, good understanding of how to train agents and how to, know, calibrate them based on getting some decisions done.

the word with partners who can execute these things along with that then evaluating on their suggestions and recommendations would be more suggestive rather than looking for tools is what I have said.

Koushik From Momentum91 (23:18)

it's

I think they can always look into our podcasts and our blogs everywhere.

Jay From Momentum91 (23:33)

Yes.

Yash From Momentum91 (23:34)

Yeah, this is one

of those things where strongly recommended that just talk to a person about this. It's helpful. another thing that might be more helpful is that you just, Koushik, you asked this question just one week earlier than I would have wanted because we're writing an e-book, right? We're writing an e-book on this very specific topic, which is role of AI agents in digital transformation.

Koushik From Momentum91 (23:40)

Yes.

Yash From Momentum91 (24:00)

And once we write that, we'll link it in the description here as well. So that should be helpful. Another thing that I wanted to understand, Jay, from an AI agent sort of standpoint is, how does a typical implementation journey look like? So let's say the first part we've covered, which is where there are four categories you try and figure out as to...

Jay From Momentum91 (24:05)

you

Yash From Momentum91 (24:23)

you know, what all processes does a business have across these four categories? And then once you have some idea, you talk to a person who can do this for you. And then let's say you arrive at three or two processes that you want to have AI agents in. Once you arrive at that, then from that to, you know, success or that to promising that same amount of people will be able to generate more value. What is that journey?

from identification, discovery of the journey that we want to fix to my people just generating like 15x more value. How does that look like?

Jay From Momentum91 (24:59)

Right. So I mean, again, this is going to be very industry specific, right? Because to give you a certain example, let's say it's a manufacturing industry doing making printers. Now, let's say they come up with a solution, come up with a question that, you know, every four years there's a downtime. And I want to, you know, improve on that. What happens is we are integrating AI in the complete

data structure, we are getting all the data points from printer, like what specific I would say, you know, equipment are being used, what it is made up from and things like that. It actually takes time for the agent to understand on, you know, like, what is the downtime, how procedures are happening, and it will also learn from what happens in these four years or the course of four years. There was one very good like key study where, know,

a company was able to utilize AI agents for just improving the overall time of overall runtime from four years to five and a half to six years. So that is like increasing almost 45 to 50 % and that was just through AI agents. So again, this was industry specific. The reason I give this example is I'm now I'll now come to the point which you asked. So once the pain points are understood, once the partner is decided how things go is

Obviously again setting up on like CDPs that is like customer data stand, customer data platforms. First of all, you need to have a very defined how things are going to be. And the very first thing will be to set up systems where data collection is happening very strongly. So that is one. Second would be integrating the right AI. So first understanding that, okay, for this particular use case, you need to integrate this set of AI. And then based on this,

you just start training the model. So the second phase would be training the model. The third would be, which will run parallel, is more about what sort of decisions you want AI models to take. And based on this, the training needs to happen. It needs to be monitored, and it needs to be altered in certain aspects as well. However, the very important and critical part, which I believe we missed out on checking on, is

There are certain limitations as well, right? And that needs to be, that's also a trending topic. And I would say needs to be covered up in this process. One thing would be biasness. So whatever model we choose based on whatever learning it has come up with, there are instances where some amount of biasness are coming through. We need to evaluate whether it is, you know, actually giving some bias in our use case or not. So that is one. Second would be the compliance side.

where security and compliance is being taken care of by implementation of this. So the third would be understanding on how and when, like how it is going to aid human beings and what part of work of humans are going to be replaced out of it. So while we are doing this process of data collection, training the AI, making sure the compliance are well set and the model is being chosen correctly and trained in that direction, also defining through service designs on

where what is the role of humans? What is the role of AI? So whatever we discussed, know, through whatever Paushy get discussed on that particular day on service design, the blueprint, there will be one more, you know, person or would say one more system that is getting added into this. And that will be replacing the part of humans. But also it will be aiding humans on getting things done faster. So that's how the whole structure would look like. Obviously the timeline. Sorry.

Obviously the timeline can vary from case to case because we are talking about lots of use cases and it is very specific. But this is how I put preliminary.

Yash From Momentum91 (28:33)

Yeah, of course.

Got it. And before we let you go, I know we are towards the end of our conversation, but before we let you go, just one last thing, because we've mentioned this recurringly almost in all answers. One of the things that's coming out is that the customer data has to be organized. And so without that, it's extremely currently it's like unorganized and it lives in the mind of the business leader or the

the person who's running the company largely is as a pattern view what we've seen. But talk about this a little, which is where, let's say if the data is organized, what sort of security features, how do I, so one of the fears that I have is that if my data is organized and is hosted somewhere, then how many people have access to it? Will it be taken away? I don't want to do that. So it is one of the fears that a lot of at least mid-market enterprises

mid-market companies have. How do we deal with that?

Jay From Momentum91 (29:37)

One interesting use case of that would be having, you know, having whole AI model in your local server. So that's what the trending topic is where you can utilize deep seek as a model. And I'm not suggesting anything because we cannot be biased on any model as of now needs to be totally use case specific. And that's where I say that understanding of right model based on our use case. if the, if the thought process is that I want to have my whole data, I want to be very secure with respect to that.

Yash From Momentum91 (29:46)

Yeah.

Of course.

Jay From Momentum91 (30:05)

and I just want it to be, know, it should not, my data should not go in the internet by any means. I can have like deep seek in local server. The model is well enough equipped in terms of getting large amount of tasks then. So we can train from that and it can be utilized. security concern is sub, it's very, I would say it is something which is.

very it is dividing world into two parts, right? The recent recent event in even France. It was all about the same where few people raise the concern on, you know, what is the security in terms of data or all over as an AI in model because many people don't know. And again, USA also mentioned that that's not the topic we are going to cover. We are only going to see the positive sides of it and then control how if there are any negative sides of it. And so

There is a part of the world which is completely open on utilizing AI in all the possible ways. And I can see from the news and things that are happening in the world, larger part of the world is going in that direction and there is no option to stay back. However, there are certain parts of the world which are also concerned about these things and still following the traditional way. However, I would say that people who have adopted the same are still secure so far and are able to

generate great amount of value out of it. So that's where we currently are.

I guess you are on mute, Yash.

Koushik From Momentum91 (31:29)

your

ideas.

Yash From Momentum91 (31:30)

correct. No, I was just saying interesting take on that. But that brings us to end of this conversation and for all the people who joined in and who are going to watch later on LinkedIn or on YouTube. Thank you for joining in. I hope this conversation was valuable for you. just before we go, whichever platform that you are on, consider

liking, subscribing to the channel. All three of us, our self-esteem sort of rides on your subscribing to the channel. We get our validation. like, you know, solving problems for customers makes us happy. But seeing the fact that you have subscribed to our channel and you like the content that we've produced makes us elated, right? Which is

which is sort of a higher form of happiness. But thank you, everyone, for joining in. And until next time, bye.

Jay From Momentum91 (32:23)

Right.

‍