Ep. 20- From Data to Decisions: Democratising AI in Marketing.

In this episode, host Ellie Tehrani sits down with AI and data analytics leader Koyelia Ghosh Roy to unpack the transformative power of AI in modern marketing. From democratising data insights to creating hyper-personalised customer experiences, Koyelia shares invaluable insights from her journey from business analyst to AI leadership. Discover how leading brands like Starbucks and Nike leverage AI, learn about the pitfalls to avoid and understand what skills marketers need to thrive in an AI-driven future. 

Transcript

00:10
Ellie Tehrani
All right. Hi, Koelia, and welcome to the Elusive Consumer. We’re very excited to have you here with us today to talk to us about all things AI and data analytics.

00:23
Koyelia Ghosh Roy
Thank you so much, Ellie. I’m really looking forward to it. I really think Elusive is doing a great job. All the podcasts that they have hosted have been insightful, data-driven, and enlightening. So, I’m really happy to be part of this podcast. I’m really looking forward to it. I’m excited.

00:43
Ellie Tehrani
Thank you so much. It’s great to have you. I’m going to jump straight into your journey from business analysts to becoming a leader in AI. Can you share what that journey looked like?

00:58
Koyelia Ghosh Roy
Oh, God. So that journey has been really quite a twist and turns, I have to say. I learned my way through my experience. So when it started as a business analyst, I got very much driven to the data, to numbers and how they talk with each other. And if you stitch together, it gives a story with data that is much more powerful than assumptions and simple visuals. So, to make a visual talk to the users in their language using data. That’s what really motivated me to move from a business analyst to a data analyst. And when I moved to data analysis, I really discovered a world of possibilities. Data can give you lot of opportunities, starting from playing with the data to transforming the data to generating meaningful insights.

01:58
Koyelia Ghosh Roy
And while we’re doing that and the plethora and mammoth amount of data that you have, you get automatically driven to the AI part, right? And when you get driven to the AI, you start really digging into deeper, you start analyzing, you start experimenting with models and with no time, you actually get so familiarise with them, even without knowing the model names that time. Because if you just know the business problem, you know the data, you know what the attributes of the data are. So you really don’t need to be absolutely a nerd and an expert in machine language, right? You don’t need to know what the data is talking about. And then you could play with these AI models and then came generative AI. And that was like a turning point.

02:49
Koyelia Ghosh Roy
The huge amount of data that we have now could actually be democratised, and it could then add insights by people who are really not that well versed with all these technical glitches or technical all kinds of algorithms, but they can still use the usage, right? And gain value out of it. That made me move into a hardcore AI field, and I started working around it with no time, actually, I’m on Into all hackathons and all kinds of POCs and getting the best in class solutions done. And it really motivates me. So, one thing I’d say that two things really helped me: my innovative mindset. Always seeking to get something new from the data. Like I used to always think differently, why it has to be done differently this way, why can’t we do a different way?

03:46
Koyelia Ghosh Roy
But that question of why it cannot be done, what helps me grow and mature into this seat, that’s my journey.

03:56
Ellie Tehrani
Wow, you talk like someone with a true passion for data, which is, it’s refreshing to hear, and also someone who understands and appreciates the importance of what data, what insights the data can give you if used appropriately. And I want to go back to the point you mentioned about democratising data insights. That’s one of your key objectives. But could you please expand on why you feel that’s important?

04:25
Koyelia Ghosh Roy
What happens is that every organisation has a mammoth amount of data, and there’s a wide range of data that they have. And with this huge amount of data comes the confusion of having sunk into the data sea. Now, without really losing the meaning of that data, how do I bring value to the end user? For example, if I have to sell a product, I will have the customer data, I will have the sales data and I will have perhaps the target customer’s profile data. Now if I don’t stitch them together and I don’t give it to the marketer, they will never be able to appreciate the data. It is just redundant data for them. Now that’s where to connect this data together, transform it, meet the needs of the users as to what they’re exactly looking for, and then give them the insights.

05:29
Koyelia Ghosh Roy
Now that’s called democratisation. Now what if I just build it for one specific group of people or one customer profile? It’s not helping. There are so many users who want to use the data but in a different way. How do I do that? Now that’s called democratisation. When the data that we have, we don’t just keep it and only a few people who know the algorithms work on it and give them, but users can use the data and derive what they want. So that’s what gives the real value.

06:01
Ellie Tehrani
And from a business perspective, you’ve touched upon this a little bit already. But in terms of how AI is transforming the way organisations understand and reach their customers, can you expand a bit on that as well?

06:19
Koyelia Ghosh Roy
Yeah, absolutely. Now we know that this digital world is continuously progressing, right? And it’s progressing at a very rapid pace. And with that comes a huge amount of your customer data, which includes your behavioral data, which includes your contextual data, the demographic data, and so much the data you have. Now. Harnessing the power of data with AI is what is going to be the key masterstroke of bringing about change. And it is no longer a luxury, it is a necessity. And at the tapestry of this vivid technology, you know what’s happening. It is the data that is there. So now the first thing that needs to be done is really transforming and using and cultivating the power of richer data.

07:14
Koyelia Ghosh Roy
So setting up data management that will break the data silos and bring the wide range of customer data that one has, including all kinds of dimensions, as I already mentioned, right? The behavioural, the demographic and others, to gain a deeper insight into your target audience. And it should not require too much data movement. There will be almost zero data movement and complete data ownership based on the strategy and based on the user group so that the marketers and the sales have control over the data and they can use it. So that’s one very specific area where data and AI merge together. And obviously with this data that you have, the second thing that you need to do is use it for analytics. And when you say analytics, it does not mean descriptive analytics like historical data.

08:16
Koyelia Ghosh Roy
You’re thinking, no, use it furthermore for predictive analytics because that will help you not just to identify the trend but actually anticipate the trend before time. So that is where the entire data activation power comes in. So one, use it for your predictive analytics. Second, use it for analyzing the trends and gaining insights. And finally, activate your data and activate it in real time so that the market can respond to the customer needs in a very contextual and relevant experience is delivered.

08:58
Ellie Tehrani
That’s a great point. And I wanted to follow up on that with regard to creating experiences, customer experiences that might actually be personalised. What role do you think the AI plays in that? Because lately there’s been a lot of talks of AI agents and AI companions that are becoming increasingly more focused on the tone of delivery beyond accurate facts. Do you think that sort of shift will help with delivering more personalised and human like customer experiences?

09:38
Koyelia Ghosh Roy
Absolutely. The first thing is that the AI enables you to first deliver insights, and based on the insights, then tailor your email campaign, tailor your ad campaigns, and study the customer journey in real-time so that you can change your advertising depending on that customer’s needs in real-time. Now that’s what makes this personalised. Every customer is unique and we need to make them feel that way. If our messaging Makes them feel like one in a crowd. We need to rethink of our approach. How do we do it? We can only do it when we study our data and then go through these entire tons of data that we have through using AI model power. That will then get us into identifying the needs of the customer, identifying the target audience and then personalising our solutions around it.

10:46
Koyelia Ghosh Roy
Now there’s a very important thing that we need to think about. I can transform brands through hyper-specialisation like never before. It can create connections that feel really intimate and unforgettable. But there is also a twist when AI missteps. It is not a missed opportunity. It can really hurt your brand trust. So the challenge here isn’t in using AI, but in using AI responsibly. AI everybody has to use. Now how you use it and how responsibly use it is a game changer. I’ll try to give you some examples and I think that will correlate to how things are happening. Starbucks. Now Starbucks started with a personalised recommendation approach. Okay. So their strategy was to develop an AI-powered recommendation system for the app. And what they choose to do is to suggest to their customers personalised coffee based on the user purchase history.

11:54
Koyelia Ghosh Roy
And that really made the experience more relevant. This small change improved customer satisfaction with things and strengthened brand loyalty. But in here, what are the key AI elements were used. Here, machine learning tailored to customer interactions was used, and this led to enhanced customer experience. If I give another example, what really worked well is Nike. Nike started something called Nike by you, right? Where they introduced AI design, AI-driven design tools. While customising the products, they allowed customers to design and personalise their shoes, thus creating a unique connection with the brand. Again it resulted in deeper customer loyalty and definitely a very deepened brand engagement. Right. Again, here it was AI which was tailored to their customer data and design algorithm. Now, now there are also cases where it really did not work well because the solution was not thought through.

13:03
Koyelia Ghosh Roy
So for example, Sephora launched a virtual artist. Okay, now what is a virtual artist? This is an augmented reality-based virtual artist which will feature for your makeup try-ons. Now, because the execution was not really good, the app struggled with the color matching, and there were glitches which frustrated the users, and this really harmed the brand’s trust. Hence there was a loss of credibility. Now, here is what the challenge was: it was on the user-face augmented reality applications that you were trying to do. So the key to starting having this customer experience, hyper Personalisation really successful is to first have a solid foundation in customer experience and insights. Beyond that, understanding the cultural nuances and the long-term trends can help the brand anticipate the shift in customer experiences and expectations.

14:03
Koyelia Ghosh Roy
Combining this with the ongoing iteration of testing and scaling will actually make it successful. The goal here is not just to create great experiences; if I have to say it is to build resilience, adaptability, and a highly sustainable solution that can thrive through any challenge. Gearing up not only for the long-term goal but also for the battles as well.

14:35
Ellie Tehrani
That’s interesting. You mentioned a few things that I want touch upon. You talk about making it successful only after planning appropriately and understanding your customers needs first before you create this personalised experience. But you also talked about the dangers of not using AI responsibly. Could you expand on that particular element?

15:03
Koyelia Ghosh Roy
Okay, now one of the things that we definitely need to understand is that when we talk of AI there are certain ethics that we need to follow. First when we are going with target audience or target customer profiling, one of the most important things is to be careful of bias. That’s a very important element to always think about. If there are biases, the solution will also get into a very difficult mode. So again, the lifeblood of AI is data. So having the data clean and having the data free of bias is extremely important. Next comes your pitching. Now if you’re pitching and you are generating content through AI and you’re just not giving it the right context, it can really harm your brand.

16:02
Koyelia Ghosh Roy
So the second responsibility usage is that how do we contextualise and personalise our messaging through the different tanners responsibly using the right contextual messages. It seems like I’m contradicting, but that’s what it is, right? In order to stay current and also be relevant, we need to keep evolving the content. The way we are generating ads, we have to continuously evolve it. One of the ways of doing it really well is through your. What we call your run, right? Retail is what they call a retail media network. The retail media network actually enables you to get the first-party customer data. You can actually marketers can give they can do the ad. That’s called programmatic ads. Right? Advertising.

16:57
Koyelia Ghosh Roy
Programmatic advertising will enable them to give whatever they want, and based on the customer detail, like the way the customer’s expectations shift, they can fine tune, they can enhance their messaging, and understand what is shifting and respond immediately to that. The first-party customer data empowers the marketers to actually shift to this huge amount of data and provide meaningful insights and give the leads, get the leads, filter the leads based on the target customer profile and then generate content around it and then convert them into the customers. The entire cycle obviously has to be AI driven because this requires powerful algorithms going through a huge amount of data in a very short period of time and contextualising it. So, as we can understand, there are three parts to it.

17:58
Koyelia Ghosh Roy
One is beta, the second is AI, and the third is the marketer. Without the marketer’s insights and experience, the solution might be very good but not successful.

18:12
Ellie Tehrani
And we’ll talk more about the skill sets needed in the future for marketers, for data analysts later in this show. But I wanted to go back to the point you made about biases and talk a little bit about data quality. You’ve mentioned in the past in some articles that the confidence in AI models is directly linked to the quality of the training data. How can organisations ensure that the data quality is there when implementing AI given the risks of diverse sources and biases?

18:48
Koyelia Ghosh Roy
That’s a very interesting question. And yeah, so there are some good guardrails that need to be followed when we go in for data. Right. The first thing is what is that data that matters? You will have a lot of redundant data. A customer might have data in multiple sources. Now first, cleaning up that data is very important. That’s one second when you’re transforming the data, what is the objective of transforming the data? Why are we doing this now? Why is very important? Based on that, the data transformation would happen, and then algorithms would run on that. Third, if we have data, how do we know that it is not biased or there’s a problem in the balance, right? And by testing data being too much, getting too much positive into things, giving the outcome rather than being, you know, giving the right outcome.

19:52
Koyelia Ghosh Roy
So how do we do that? How do we change that? So there, all the time, doing feature engineering and dropping the outliers is extremely important so that the data is not skewed. The data has to be normalised to initiate that analysis. And finally, once we have done that, there’s one thing that we need to understand, and we have to really see to that is one, if there are any demographic-based data, how is my data segmented? I should have equal data points across all my segmentations. So, data segmentation is also a very important parameter to identify how this is going to be read and how the algorithm is going to work on it. And finally, if I am really getting the data, Do I have permission to use the data?

20:48
Koyelia Ghosh Roy
Getting those permissions from the customers, making the customers aware that this is what we are using it for, is extremely important before you even start using your AI models or in it, and last, using it for the benefit of society, not for only the profit, but for the benefit of the society. And finally, making it sustainable. Algorithms of AI can really consume a lot of energy. How do I make it optimised so that for certain algorithms and certain transformations, I really don’t need such powerful models? So I can really reduce my GPUs to say, edge devices. I can then compress my models through the quantisation method. I can then run my models, perhaps in batch mode at times. So all this optimisation technique has to be applied so that the solution is also sustainable.

21:47
Koyelia Ghosh Roy
So these are a few points that are coming to my mind right now.

21:50
Ellie Tehrani
Yeah, there are a lot of interesting points there, including the segmentation and ensuring that there is an appropriate representation of your audience and not just skewed towards some selection. But I also want to talk to you a little bit more about the cleaning that you mentioned. There’s often this concept of garbage in, garbage out. In discussions around AI, are there particular processes or advice that you recommend for data cleaning and data validation to ensure that it is less garbage coming in, so to speak?

22:28
Koyelia Ghosh Roy
So as a. Okay, so there are many very technical ways of doing it. Okay, so one, as I said, outliers, right? You first normalise the data and then identify the outliers, and then you clean the data and go around and do it. Right? The other One is the P1 score, F1 score, right? You need to really see, okay, if I’m out of the all the data, what is my score with respect to my data distribution and my parameters against that data distribution? Okay, if they’re equally distributed, then Third is from a very business perspective, if I say the attributes of my data, which is called metadata, also what is the data source? What does it talk about? What does it imply? What was the date it was updated? What is the source of it?

23:16
Koyelia Ghosh Roy
Those are very specific information that one needs to understand in order to clean the data. Now, suppose you are mixing 2023 data with 2024. Your data will obviously be skewed, right? Your predictions will all be incorrect. The third one is, of course, on the seasonality. Now, if you click a data and from a sales perspective, there’s a seasonality which impacts, and you just use the entire thing to do predictive analytics. The entire data and your entire results would be skewed. So when we clean the data, we need also to see this parameter whether, okay, what are the different seasonality data that I need to normalise then, or I have to make them as drop them from my analysis at all, or I have only to take the seasonal data and do a forecast analysis around it.

24:04
Koyelia Ghosh Roy
Then there’s of course, as I said, feature engineering, like which features I want to bring in and which features I want to study in this analysis. So in that I will only pick up those features and I will read through them and I’ll then perform manas. Did I answer your question or did we get too technical?

24:19
Ellie Tehrani
No, absolutely. I think this is all very helpful to understand what processes are involved, to ensure that during that cleaning, you’re picking up as much of that as possible. But to another point about the data sources, because companies – or – particularly larger enterprises, also have access to their own data sets. So when it comes to balancing the need to rely on external sources, which is often more costly, versus building your own, how do you think organisations should decide there? And when should they rely on external sources versus internal data sets?

25:02
Koyelia Ghosh Roy
When we talk about internal data sets, that helps in two things. One is understanding the data. One needs to really understand the data to really perform that. And that comes with our historical data. That gives us a lot of confidence that we now understand, helps us understand the customer insights, helps us understand the trends, and helps us bring about meaningful data insights from that data set. So that’s one. But where the market shifts, the historical data will never give you the idea that the market is shifting. That’s where real-time data from external sources are required. Now, it is always advisable that we should go with reliable data sources like some analysts or some media houses which are reliable, to provide reliable data sources.

25:56
Koyelia Ghosh Roy
Because if we go with Internet data, that will be cartridge and coverage out so that level of maturity only very few firms would bring. Right. That’s the reason why different firms are really relied on that. So that maturity of which data to use for your analysis with respect to the real time data and with the market shift data. Okay, so it could be across a period that we take it from the external, and then we merge our trend of historical trend data to see how the factors are impacting each other, which are the inverse of the factor related and directly related. Right. So that we can then bring those factors and then we can attach it to the attributes of the data and then give analysis around it. Now that we have the data not necessarily.

26:51
Koyelia Ghosh Roy
We’ll have that amount of data for which a model needs to be trained on. Now, that’s where we generate data, right? Synthetic data. That synthetic data then comes into use in such cases where we identify the trend, we see the attributes and we then generate synthetic data around it and train our model to that. Our training data set doesn’t need a huge amount of the current data that is coming, the data that is real-time flowing in. Right. And then when we have to test, we can then test with the flowing data so that we know the hypothesis. Correct.

27:29
Ellie Tehrani
That’s excellent. I think that the second question, the follow up on that is let’s say you have used AI models and you have done your research, you’ve gathered this data, you have the appropriate skill sets in house to review and tell the story with that data. But how do you now ensure that the insights is actually used to drive business decisions?

27:59
Koyelia Ghosh Roy
Absolutely. Now, that’s a very important point, and we have seen in a number of areas that the insights, although they are getting the insight, are not really being used. Right. So there are different ways where these insights can actually be used. As I said, the true power lies in using these insights and then using AI on this insight to generate value. So an example I have to say is that if you have an insight and now you know what your customer profile is, the target customer profile is, you know, what is their customer expectation. Now, you need to do automated email marketing. AI can help you cater to personal customers with a very personalised customer experience by automating this email. It analyzes the user behaviour and then crafts personalised messages. This enables real engagements with the customer and then leads to conversion.

29:03
Koyelia Ghosh Roy
So that’s one of the ways in which the insights are converted into value. The other one is about your market campaign design. Again, with Genai, they can actually design various versions of the marketing campaigns, which is time-consuming currently, if you have to say. But with AI and with Genai having this multimodal power, it can help you generate different versions in a few seconds and then they can actually pick and forecast, you know, what is that get them the potential taxes. Then there’s also this point of your zero-click search strategy. So what happens is that if you are working, if the business is working on getting a prominent spot in your search result page, right? Search engine result page. Now, no user will actually click on the links.

29:57
Koyelia Ghosh Roy
How does the business really create a very engaging snippet of the offering that comes directly onto the SERP so that users don’t need to click on that link, they immediately see that and the moment they see it, that message and snippet worthy messaging, which is so direct and relevant based on the customer journey that they had that they will definitely get converted. Now, that is a very good utilisation of insights and getting converted into personalised messages, which are then directly going into the SERP and having a zero-click search strategy. So immediately you have visibility, your visibility dramatically increases. The second way also is that you can optimise your what we can see AB testing.

30:50
Koyelia Ghosh Roy
So what is B testing is that you need to create a lot of variations to understand you know which one really works and that AI actually helps you to really do that testing. So that is another way of actually enabling and getting your solution, your campaigns and your email generation ad creations. Really much more effective.

31:15
Ellie Tehrani
That’s great. I think that’s been one of the main issues generally across many businesses that sometimes it’s not so much about lacking the data or even the knowledge of how to interpret the data, but rather how to convince the right stakeholders of the best ways of using that data to implement correct business decisions. So that’s very helpful. Thank you. I want to move on quickly to some future trends and challenges. What do you see being the biggest challenge for marketing in this AI-driven future?

31:55
Koyelia Ghosh Roy
Okay, now the challenge that they can have in this is I would again emphasise on this one thing, activating data. Now, what do we mean? Activating data is getting the right data at the right time and utilising it. The biggest problem here is that data is there, but it’s not getting utilised or that is not getting placed at the right time. So, these are the two areas that marketers need to work on. They really need to understand how they actually activate the data. The activating of the data could be based on the customer insights that they have in the real-time because customer expectations are shifting very rapidly. So quickly getting that data access, generating insights, and then utilising it, as I just mentioned, right?

33:03
Koyelia Ghosh Roy
Filtering through the web, getting the target, getting the leads, filtering the leads based on the customer profile and then generating content and then converting your leads into customers. Now, that entire cycle requires data activation. Now, that’s a challenge for marketers. The one who can do it appropriately will be able to gain success in a much shorter time. So that’s one challenge that I do see that marketers need to navigate through very intelligently.

33:36
Ellie Tehrani
And that’s one challenge with finding also the appropriate marketers for your organisation or upskilling your current team. But what other challenges are there? Or how can rather organisations better prepare for this increasing role of AI in customer interactions?

33:55
Koyelia Ghosh Roy
One, of course, as I said that customer data is extremely important to get customer insights, right? And then get from the customer insights then drive your product development. So based on that, the way the marketers actually work, right? So they are actually continuously trying to reach their customers, their target customers, right? So they really need to overhaul the traditional customer behavior analysis model that they’re doing. The AI has the power to connect data points, they can identify patterns, they can learn from them, and they can use information to give you a plethora of analysis and customer behavior. But it is the power of the marketer that they need to understand what is the segment that they are targeting, what is the product that the customer would be really interested in? What are the hyper-personalisation criteria that can convert that lead into the customer?

35:01
Koyelia Ghosh Roy
Now those are the areas where the marketer needs to work. But then they are not alone there. The AI will actually also help them to identify, okay, this is what the shifting priorities are. This is where the consumer trends going and what are the four or five priorities it thinks the customer would be interested in. So it’s really narrowing down the focus of the marketers with earlier the marketer had to look through a wider range of data. Now it’s a very narrow data. So the more narrow the data is, the more focused the data is, and the more insights we can focus on. Their pain point is that it helps them to actually derive products based on customer insights.

35:45
Ellie Tehrani
That’s interesting that the tools actually help them focus better. So that brings me to the question about what are the most important skills for marketers to learn in the next few years, do you think?

35:58
Koyelia Ghosh Roy
One of the things that I would say is that they really need to understand data attributes. You know, what makes the data really worthy, really valuable? Okay. They end, they really don’t need to be technically really efficient or really great in that they need to use their power of understanding a customer. And that has sharpened that. What attributes I need, what are my metadata or what I mean by metadata is what are the specific criteria of my data that we need to do our analysis on. Then it goes into the next step of how do I connect with my customers, what are my platforms I’m using, as I said recently, the rmn, the retail media network that are happening, they’re really helping the marketers to connect, right? So understanding these new platforms that how do they work, they will really help them.

36:57
Koyelia Ghosh Roy
The second Is getting used to this granite solution, which doesn’t need them to really get into the back end of it where there are a lot of algorithms working behind it. But knowing what they exactly want from them really be good in their prompt engineering techniques. Okay. It is very important to frame their problems so that they get the maximum benefit and output from that model as per their need. So, three things. Understanding your data better helps you identify the attributes. The second is prompt engineering. Third what are the activities you want the AI to do? Do you want the AI to do a search for you and bring relevant results? Do you want the relevant results to be tacked differently? That’s what NSA understands, and they need the AI to do. That’s the AI agentic workflow I’m talking about.

38:01
Koyelia Ghosh Roy
Because AI agents can be easily be done. And with this advancement of AI, you can actually just interact and easily do it in a low-code, no-code fashion. But what you really want that to be done, what that agent should do that is important. And when we talked about democratising AI, I actually meant that now you, I, anybody can really become a very good professional who is working with AI. Earlier, that was not the case. Right. So now this has come to that level.

38:39
Ellie Tehrani
Absolutely. It’s allowing all of us to become more informed and use that data to improve ourselves and also better support our organisations. But from the organisation’s perspective, if you have a business that is looking to become more data-informed, what advice would you give to them?

39:03
Koyelia Ghosh Roy
I would give them two things. You know. First, harness your data. Harness the power of your data. Harnessing the power of data means cleaning the data, understanding the data and identifying what will generate value for you. Then, identifying the sources from external sites, what would give them the benefit? Once that is known, the rest is very easy. If data is the fuel of AI, right? If the data is right, your AI will also be right. So having the data correct, having the data really activated is what is important.

39:44
Ellie Tehrani
And finally, as we have a few more minutes left in this show, are there any final thoughts that we haven’t discussed that you’d like to share with our audience about the future of data and AI in marketing?

40:00
Koyelia Ghosh Roy
I think we talked about this consumer behaviour analysis with AI. Right? We also talked about hyper-personalisation, and I think we also talked about bridging the gap between the data insights and the execution. Right? I mean how do we really use the insights? I would lastly say that innovations come from your experience, and one can enhance their marketing efforts by leveraging the power of data and filtering them based on their needs. So that’s what I would request all the marketers to do. Keep innovating by harnessing the power of data.

40:47
Ellie Tehrani
Brilliant advice. Thank you so much, Koila, for your time today.

About Our Guest