Snowflake Data Cloud Summit 2024 | Cindi Howson | ThoughtSpot & Mark D’Costa | BNY Mellon

Clips
News
More from Snowflake Data Cloud Summit 2024

Cindi Howson

Chief Data Strategy Officer

ThoughtSpot

Mark D'Costa

Head of Data Innovation & Commercialization

BNY Mellon

BNY Mellon pioneers AI data integration thanks to ThoughtSpot and Snowflake partnership

To add fuel to the artificial intelligence fire, data ought to take center stage, and this is why AI data integration has become top of mind across different industries. To incorporate AI capabilities into its data estate, the Bank of New York Mellon Corp. teamed up with ThoughtSpot Inc. and Snowflake Inc., according to Mark D’Costa (pictured, right), head of data innovation and commercialization at BNY Mellon.ThoughtSpot’s Cindi Howson and BNY Mellon’s Mark D’Costa talk to theCUBE about the importance of AI data integration in banking.“The LLMs and generative AI have just been an amazing kind of jolt to the organization in the industry,” D’Costa said

play_circle_outline Introduction to the Snowflake Data Cloud Summit and hosts Rebecca Knight and Dave Vellante

play_circle_outline The Transformation of the Chief Data Officer Role: Insights from Mark D'Costa and Cindi Howson

play_circle_outline The need for organizations to disrupt themselves and stay competitive

play_circle_outline Benefits of using Snowflake and ThoughtSpot for data innovation and commercialization

play_circle_outline Aspirational goals for the future, including a governed, tenanted Snowflake platform and expanding ThoughtSpot's SaaS offering.

Info

Cindi Howson | ThoughtSpot & Mark D’Costa | BNY Mellon

Cindi Howson

Chief Data Strategy Officer ThoughtSpot

Mark D'Costa

Head of Data Innovation & Commercialization BNY Mellon

search

Rebecca Knight

>> Good morning everyone, and welcome back to theCUBE's Live coverage of the Snowflake Data Cloud Summit. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Dave Vellante. Dave, the keynotes of the morning are finished, the buzz and energy on the show floor is really exciting. There's lots of excitement going.

>> It's still buzzing. I mean, last night was a lot of action and it's still packed today. I mean, really packed.

Rebecca Knight

>> Yeah, day three of the conference, but things are really heating up. I'd like to introduce our next two guests. We have Mark D'Costa. He is the head of Data Innovation and Commercialization at BNY Mellon. Welcome, Mark.

Mark D'Costa

>> Thank you.

Rebecca Knight

>> And Cindi Howson, chief data strategy officer at ThoughtSpot. Thank you so much for coming on theCUBE, both of you.

Cindi Howson

>> Thanks for having both of us.

Rebecca Knight

>> So both of you have really cool titles, particularly for this moment in time in the enterprise AI era. Mark, I want to start with you. Tell us a little bit about your role at BNY Mellon.

Mark D'Costa

>> So I report to the Chief Data Officer at BNY Mellon, and my role is head of data innovation. And what is that? It's really about data strategy and innovation. And what's really great about my role is I get to look at some of the latest things that are coming down the pike, assess it from a strategy perspective, and in many instances take that and then hand it off to other teams. One of the things I've been able to do is with ThoughtSpot and then Snowflake, those are points where I innovated them, worked on the strategy and then never let them go because I really think that those are going to be really transformative pieces to both the company's future and their clients.

Rebecca Knight

>> Cindi, as you know, the chief data officer role has involved quite dramatically. It used to be this back office role, it wasn't even called the chief data officer, but it was this data quality, keep it in the back room and now firms like ThoughtSpot have brought it to the fore so that end users can actually interact with data. Of course, you still have to trust the data. How have you seen that role evolve in your career?

Cindi Howson

>> And this is where we're even seeing a change in the title to chief data and analytics officer, and in some organizations it's chief data and AI officer, because if you're only collecting data and not putting it in the hands of the business users or at the point of impact, you're really just a cost. And it's about going from that defensive, protect the data, to offensive, make that data actionable and get the insights.

>> So financial services, so the CDO role first emerged really in regulated industries, financial services, healthcare, public sector, and in a way you should be in theory in a better position to take advantage of AI because everybody's really concerned about data quality and the legal aspects of it. How do you feel about that in terms of the bank's ability to add AI to its data estate? What's the state of data?

Mark D'Costa

>> So the LLMs and generative AI has just been an amazing kind of jolt to the organization in the industry. The focus on that new technology and the prospect that it has, has then just turbocharge the focus on data because you need to have that data in order to power all of those new cool things. And so it's been really complimentary to what we've been doing within the chief data office. We've been there focused on data largely for driving, to your point, regulatory items. Now that same data, that same governed data is there to power commercial outcomes. So the timing in terms of being in this space and being in the chief data office space at the time where AI and LLM is just really taking the hill has been really an exciting place to be.

>> So Cindi, the old saying, "Why do you rob banks?" Well, that's where the money is. Well, the money's in financial services, everybody wants to disrupt financial services. How do you see GenAI affecting financial services? Of course, you had the whole crypto thing, you've got changes in wealth management. Is it ripe for disruption? And how do you help customers like Mark not get disrupted or disrupt themselves?

Cindi Howson

>> Yeah, no, they have to disrupt themselves. That's why Mark also is charged with innovation. There's keeping the lights on, but there's also innovation. And you made an important distinction in your part two question, Dave, and we need your help with this. There's a lot of AI washing and we have to distinguish between GenAI, which has given the industry, the jolt that Mark referred to, and regular AI machine learning that of course, if I apply for a mortgage at BNY, they're using traditional AI. The GenAI is what is really important and revitalizing this industry that has taken it now to a boardroom conversation. What is our data strategy? Because without a data strategy, you cannot have an AI strategy. And how quickly can we take advantage of GenAI. With BNY Mellon investing in products and platforms like Snowflake that puts them in a better position to leverage generative AI?

Rebecca Knight

>> Well, this is really interesting because it's about organizational change here. It's really about shifting mindsets, changing the way people think about, but also how they do their jobs. What are you seeing in your organizations in terms of bringing people along? Because as you said, it's a boardroom conversation, but that has to filter down to the rank and file too.

Mark D'Costa

>> So I mean, if it's all right, I'll take that first. So it's all about timing. So as this whole focus on generative AI and data transformation is taking place, the organization's also going through an organizational transformation. So we're going through a platform operating model. So think about this in terms of previously we grew up and we might have had payments sitting across 18 different lines of business, now all of that is being aligned into one product for us to go to market with clients. Those teams in turn are turning around looking at their data infrastructure and strategy and realizing, I need more. I need scale, I need cloud scale. So's there's a big element of being in the right place at the right time. And so what we had been doing with Snowflake in terms of creating a very tenanted structure that was governed to enable all these different lines of business that came together at the right time, we're doing the same thing with ThoughtSpot. As we start to shift the data gravity to the cloud, we're trying to offer ThoughtSpot as a tenanted service as well. Once we get to cloud, we can start to let off the reins a bit more and allow the teams to individually manage these tenanted pieces of infrastructure and services in order to meet the needs for their businesses and their clients.

Rebecca Knight

>> Explain why. Is it just a flexibility? Is there a cost factor, a scalability?

Mark D'Costa

>> I think it's a bit of both. So we've been on the Snowflake journey in earnest for about a year in the regulated entity, we've been on the ThoughtSpot journey for three years. So for the first two years we are working with ThoughtSpot, which is a cloud powered solution, but in an on-prem mode. Very constrained, you can try to get it to work... We got it to work, we got great outcomes, but you're ultimately running in against hardware limitations and scale. So as we started to break through into the cloud with Snowflake, you're now no longer having to shift and move and copy data. ThoughtSpot has a really thin layer that compliments Snowflake very well, it pushes down all that processing to Snowflake. It's not moving and shuttling the data or caching the data. It's very, very thin. And that in itself then means that you can just tap into Snowflake scale. So what you couldn't do in any way on-Prem now becomes possible once you've kind of broken through into the cloud with both ThoughtSpot and Snowflake.

>> So the cloud is the underlying infrastructure, could you talk to the business value that you're getting? Because ThoughtSpot is all about democratizing data, providing wider access. How are you taking advantage of that vision?

Mark D'Costa

>> So a lot of why we've gotten through to the cloud has also been our clients. Clients are there. They're asking, saying, "Hey, when are you going to come along?" So we're on this journey with our clients. And so what we're finding is the clients want us to meet them there. They want to start to receive their data in a connected form. They don't really want to get FTPs and files and all these different bits kind of cut and sent across to them, and they have to stitch it together. They're looking for this connected space, so we're early on this journey, but we already are partnering with clients in order to work through Snowflake sharing in order to get data just showing up on their doorstep as opposed to shuttling it across and having them reconstituted in their environment. So it's very transformational.

>> And it's live data, essentially.

Mark D'Costa

>> Live data.

>> Okay. And so Cindi, I want to ask you, because of your experience as an analyst prior to ThoughtSpot, and you've seen the pendulum swings. I won't go back to the mainframe and the early Oracle days-

Cindi Howson

>> .

>> But I will go back to Hadoop. The promise of Hadoop was, hey, we're going to solve all the world's problems. We're going to bring the compute to the data. And then we found out it was way too hard, and then you had to be sort of a geek to really get into Spark land, that's fine. Snowflake comes along and says, "Hey, we're going to make this really easy." Great, so put everything into Snowflake, run ThoughtSpot on top, scale of the cloud, accessibility. Now, this is all these open data formats, so the pendulum is now swinging back. How do you ensure that there is, I know it's an overused term, but this single version of the truth, that billings means the same thing no matter which application it's in. What's your experience there and what's your vision in terms of preserving that unity?

Cindi Howson

>> So as Mark mentioned, it's about not moving the data. That is a legacy process, and Rebecca, you referred to the people change management, and so moving the data is a legacy process and mindset. Setting up things like data sharing and clean rooms is really important. And I come back to the data is just one piece of the puzzle, that things that Bank of New York Mellon has accomplished with ThoughtSpot, letting the business users ask their questions at the point of impact without having to go through a dashboard developer as an expert and as a middle person, is really what changes the game and unlocks the value of all that data. So whether it's, and we can't get too specific, whether it's, what are payment times looking for this one line of business versus another and asking that in natural language, this is also where the people change management, there are power shifts. So when you have democratized data access, those that are performing well, love it. Those who are not performing well, now you have to say, do we have a good culture, a learning culture? And just another industry, a CPG customer that I work with through this, they were able to identify stockouts and eliminate stockouts by 20%. That is millions of dollars in potentially missed sales from a stockout. So I think this power shift along with the tech modernization and democratized insights is where the true value is.

Rebecca Knight

>> So is it changing the way organizations recruit? I mean, what you're talking about is this learning culture, this growth mindset, this having to learn new technologies and as you said, the people who are doing well continue to do well because that's what they're doing already and the laggards fall by the wayside. But how would you describe how it's changing the way organizations, the kind of skills that they're looking for and the kind of traits and attributes they want in employees?

Cindi Howson

>> Yeah. And so I'll speak industry level and then Mark, you can share from your team, the talent wars have never been fiercer. Everyone wants to hire an LLM expert today. But you know what? LLM experts that existed were in the handful 18 months ago. So what you need to recruit for is curiosity, and who is the fastest learner that is both from the core data job as well as the leaders. The leaders that can combine business, what's their imagination? How can I apply generative AI to this business problem? And they understand the tech stack,. Those are the ones that are much further ahead in this wave of disruption than those that stick their head in the sand and wake me up when this wave is over.

Mark D'Costa

>> What I'm also seeing though, is that the power of some of the generative AI solutions that we have, like within our company there's a AI hub team that's really focused on taking that hill as it relates to AI within the firm and the strategy and they have an app that they've wrapped around these tools called Eliza. And what I've seen with that is that it really brings the access barrier way down. So in a number of instances, you don't necessarily need lots of the specialists and the LLM modelers, what you need are people who are creative, understand their business flow, and have been asking for tools and methods in order to solve problems. So what I'm really seeing is that there's a lot of energy within the firm because there are teams that know their businesses, know their clients really well, and now they suddenly have this additional, let's call it AI assist copilot in order to help them bring those ideas to fruition.

Cindi Howson

>> Eliza, that's an interesting name. Do you know-

Mark D'Costa

>> It is.

>> I wonder if it's the same? Did you know there was a... The first chatbot I believe ever was called Eliza.

Mark D'Costa

>> There's a story around Eliza. I'm not going to try to say it because I'll get it wrong, but think it has some tie ins with our firm's history and some other bits.

Cindi Howson

>> So yeah, they called it a chatter bot back then, and they were amazed like we are today at GenAI, and then it sort of died.

Mark D'Costa

>> Eliza's pretty smart now and she's getting smarter by the day exponentially.

Rebecca Knight

>> Well, I mean, think about Pygmalion, Eliza Doolittle, she learns.

Cindi Howson

>> Okay, now you're taking back to my starring role in junior high.

>> Okay, I want to ask more about disruption. We had a pharma on yesterday, and this individual predicted that there'll be a 30 to 40% disruption in his data pipeline in the next three or four years. Zhamak Dehghani has done a very good job of describing the hyper specialized roles within the data pipeline and the pain of actually getting to that stock out example, you just need so many experts. How do you see Gen AI changing all the heavy lifting that we need to do?

Cindi Howson

>> Yeah. And so two parts, so Zhamak Dehghani, love her work, have had her on The Data Chief podcast as well, but I think more than the tech, what she got people to think differently about that BNY also applies is that domain ownership and agility, that you cannot just have one centralized data warehouse, slow monolithic, all the pipelines going there. So I do think gen AI will help with things like data observability, but also crowdsourcing the business definitions, do you call something bookings or revenues? So crowdsourcing the business definitions both for the pipelines and also for the dashboards, the insights that you're creating, improving data quality as well, looking for those outliers without it having to be so rule-based, those those are the opportunities I see.

Mark D'Costa

>> I also see it. I've seen some really interesting things where it will also lower the barrier to being able to work with data. So I'm a SQL guy and I can do some pretty interesting things with SQL. What I'm also seeing is that with the advent of LLMs and additional tooling, that starts to go away. And so you can start to then have people who haven't had to spend the time learning that technical skill, suddenly they're able to still work with the data, get to insights that they just couldn't otherwise get to before having lean on someone else. So that efficiency is going to be really strong from an outcome perspective. And then within the data pipelines, I do see the ability for much more AI assist in terms of classification labeling, think information classification. Think about 20,000 tables with X number of fields, the surface area is huge. So you need to have some form of human assist there in order to get it right and to scale that out, because otherwise, previously you could only govern a small amount of highly governed regulatory data, now you really want to be able to unleash that data for all sorts of outcomes. You need to have something that's going to help you scale.

Mark D'Costa

>> And I'm not a SQL guy and I don't want to be, but I know my data better than the technical people know the data. And so to the point you were making earlier, Cindi, about the organizational aspects of Zhamak's work, to be able to truly democratize the data you need that human input of those who actually are in the know. And at the same time, it's got to be governed. It can't be the wild west.

Mark D'Costa

>> It does, it does.

Rebecca Knight

>> It's got to be the curious creative liberal arts graduates. Finally, Mark, the year's half over, but I'd be interested to hear your aspirational goals for the rest of 2024 and even into 2025.

Mark D'Costa

>> So look, the key goal for us has been trying to stand up this governed, tenanted Snowflake platform that kind of unleashes creativity within the firm. Also, now adding to that, how do we make a service like ThoughtSpot fully tenanted as well? The data gravity starting to shift dramatically from on-prem to cloud. So as we get to that data gravity shifting, one of the aspirational goals is to start to look at, how do we simplify our lives? How do we start to look at ThoughtSpot's SaaS offering? How do we then... That in itself is going to create more efficiencies and unlock even more features and functionalities that we wouldn't otherwise be able to go after.

Rebecca Knight

>> Excellent. Well, Mark and Cindi both, thank you so much for coming on theCUBE, a really great conversation.

Mark D'Costa

>> Thank you.

>> Thanks guys.

Cindi Howson

>> Thank you for having us.

Rebecca Knight

>> I'm Rebecca Knight, stay tuned for more of theCUBE's live coverage of the Snowflake Data Cloud Summit. You're watching theCUBE, the leader in enterprise tech news and analysis.

Cindi Howson | ThoughtSpot & Mark D’Costa | BNY Mellon

search

Rebecca Knight

>> Good morning everyone, and welcome back to theCUBE's Live coverage of the Snowflake Data Cloud Summit. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Dave Vellante. Dave, the keynotes of the morning are finished, the buzz and energy on the show floor is really exciting. There's lots of excitement going.

>> It's still buzzing. I mean, last night was a lot of action and it's still packed today. I mean, really packed.

Rebecca Knight

>> Yeah, day three of the conference, but things are really heating up. I'd like to introduce our next two guests. We have Mark D'Costa. He is the head of Data Innovation and Commercialization at BNY Mellon. Welcome, Mark.

Mark D'Costa

>> Thank you.

Rebecca Knight

>> And Cindi Howson, chief data strategy officer at ThoughtSpot. Thank you so much for coming on theCUBE, both of you.

Cindi Howson

>> Thanks for having both of us.

Rebecca Knight

>> So both of you have really cool titles, particularly for this moment in time in the enterprise AI era. Mark, I want to start with you. Tell us a little bit about your role at BNY Mellon.

Mark D'Costa

>> So I report to the Chief Data Officer at BNY Mellon, and my role is head of data innovation. And what is that? It's really about data strategy and innovation. And what's really great about my role is I get to look at some of the latest things that are coming down the pike, assess it from a strategy perspective, and in many instances take that and then hand it off to other teams. One of the things I've been able to do is with ThoughtSpot and then Snowflake, those are points where I innovated them, worked on the strategy and then never let them go because I really think that those are going to be really transformative pieces to both the company's future and their clients.

Rebecca Knight

>> Cindi, as you know, the chief data officer role has involved quite dramatically. It used to be this back office role, it wasn't even called the chief data officer, but it was this data quality, keep it in the back room and now firms like ThoughtSpot have brought it to the fore so that end users can actually interact with data. Of course, you still have to trust the data. How have you seen that role evolve in your career?

Cindi Howson

>> And this is where we're even seeing a change in the title to chief data and analytics officer, and in some organizations it's chief data and AI officer, because if you're only collecting data and not putting it in the hands of the business users or at the point of impact, you're really just a cost. And it's about going from that defensive, protect the data, to offensive, make that data actionable and get the insights.

>> So financial services, so the CDO role first emerged really in regulated industries, financial services, healthcare, public sector, and in a way you should be in theory in a better position to take advantage of AI because everybody's really concerned about data quality and the legal aspects of it. How do you feel about that in terms of the bank's ability to add AI to its data estate? What's the state of data?

Mark D'Costa

>> So the LLMs and generative AI has just been an amazing kind of jolt to the organization in the industry. The focus on that new technology and the prospect that it has, has then just turbocharge the focus on data because you need to have that data in order to power all of those new cool things. And so it's been really complimentary to what we've been doing within the chief data office. We've been there focused on data largely for driving, to your point, regulatory items. Now that same data, that same governed data is there to power commercial outcomes. So the timing in terms of being in this space and being in the chief data office space at the time where AI and LLM is just really taking the hill has been really an exciting place to be.

>> So Cindi, the old saying, "Why do you rob banks?" Well, that's where the money is. Well, the money's in financial services, everybody wants to disrupt financial services. How do you see GenAI affecting financial services? Of course, you had the whole crypto thing, you've got changes in wealth management. Is it ripe for disruption? And how do you help customers like Mark not get disrupted or disrupt themselves?

Cindi Howson

>> Yeah, no, they have to disrupt themselves. That's why Mark also is charged with innovation. There's keeping the lights on, but there's also innovation. And you made an important distinction in your part two question, Dave, and we need your help with this. There's a lot of AI washing and we have to distinguish between GenAI, which has given the industry, the jolt that Mark referred to, and regular AI machine learning that of course, if I apply for a mortgage at BNY, they're using traditional AI. The GenAI is what is really important and revitalizing this industry that has taken it now to a boardroom conversation. What is our data strategy? Because without a data strategy, you cannot have an AI strategy. And how quickly can we take advantage of GenAI. With BNY Mellon investing in products and platforms like Snowflake that puts them in a better position to leverage generative AI?

Rebecca Knight

>> Well, this is really interesting because it's about organizational change here. It's really about shifting mindsets, changing the way people think about, but also how they do their jobs. What are you seeing in your organizations in terms of bringing people along? Because as you said, it's a boardroom conversation, but that has to filter down to the rank and file too.

Mark D'Costa

>> So I mean, if it's all right, I'll take that first. So it's all about timing. So as this whole focus on generative AI and data transformation is taking place, the organization's also going through an organizational transformation. So we're going through a platform operating model. So think about this in terms of previously we grew up and we might have had payments sitting across 18 different lines of business, now all of that is being aligned into one product for us to go to market with clients. Those teams in turn are turning around looking at their data infrastructure and strategy and realizing, I need more. I need scale, I need cloud scale. So's there's a big element of being in the right place at the right time. And so what we had been doing with Snowflake in terms of creating a very tenanted structure that was governed to enable all these different lines of business that came together at the right time, we're doing the same thing with ThoughtSpot. As we start to shift the data gravity to the cloud, we're trying to offer ThoughtSpot as a tenanted service as well. Once we get to cloud, we can start to let off the reins a bit more and allow the teams to individually manage these tenanted pieces of infrastructure and services in order to meet the needs for their businesses and their clients.

Rebecca Knight

>> Explain why. Is it just a flexibility? Is there a cost factor, a scalability?

Mark D'Costa

>> I think it's a bit of both. So we've been on the Snowflake journey in earnest for about a year in the regulated entity, we've been on the ThoughtSpot journey for three years. So for the first two years we are working with ThoughtSpot, which is a cloud powered solution, but in an on-prem mode. Very constrained, you can try to get it to work... We got it to work, we got great outcomes, but you're ultimately running in against hardware limitations and scale. So as we started to break through into the cloud with Snowflake, you're now no longer having to shift and move and copy data. ThoughtSpot has a really thin layer that compliments Snowflake very well, it pushes down all that processing to Snowflake. It's not moving and shuttling the data or caching the data. It's very, very thin. And that in itself then means that you can just tap into Snowflake scale. So what you couldn't do in any way on-Prem now becomes possible once you've kind of broken through into the cloud with both ThoughtSpot and Snowflake.

>> So the cloud is the underlying infrastructure, could you talk to the business value that you're getting? Because ThoughtSpot is all about democratizing data, providing wider access. How are you taking advantage of that vision?

Mark D'Costa

>> So a lot of why we've gotten through to the cloud has also been our clients. Clients are there. They're asking, saying, "Hey, when are you going to come along?" So we're on this journey with our clients. And so what we're finding is the clients want us to meet them there. They want to start to receive their data in a connected form. They don't really want to get FTPs and files and all these different bits kind of cut and sent across to them, and they have to stitch it together. They're looking for this connected space, so we're early on this journey, but we already are partnering with clients in order to work through Snowflake sharing in order to get data just showing up on their doorstep as opposed to shuttling it across and having them reconstituted in their environment. So it's very transformational.

>> And it's live data, essentially.

Mark D'Costa

>> Live data.

>> Okay. And so Cindi, I want to ask you, because of your experience as an analyst prior to ThoughtSpot, and you've seen the pendulum swings. I won't go back to the mainframe and the early Oracle days-

Cindi Howson

>> .

>> But I will go back to Hadoop. The promise of Hadoop was, hey, we're going to solve all the world's problems. We're going to bring the compute to the data. And then we found out it was way too hard, and then you had to be sort of a geek to really get into Spark land, that's fine. Snowflake comes along and says, "Hey, we're going to make this really easy." Great, so put everything into Snowflake, run ThoughtSpot on top, scale of the cloud, accessibility. Now, this is all these open data formats, so the pendulum is now swinging back. How do you ensure that there is, I know it's an overused term, but this single version of the truth, that billings means the same thing no matter which application it's in. What's your experience there and what's your vision in terms of preserving that unity?

Cindi Howson

>> So as Mark mentioned, it's about not moving the data. That is a legacy process, and Rebecca, you referred to the people change management, and so moving the data is a legacy process and mindset. Setting up things like data sharing and clean rooms is really important. And I come back to the data is just one piece of the puzzle, that things that Bank of New York Mellon has accomplished with ThoughtSpot, letting the business users ask their questions at the point of impact without having to go through a dashboard developer as an expert and as a middle person, is really what changes the game and unlocks the value of all that data. So whether it's, and we can't get too specific, whether it's, what are payment times looking for this one line of business versus another and asking that in natural language, this is also where the people change management, there are power shifts. So when you have democratized data access, those that are performing well, love it. Those who are not performing well, now you have to say, do we have a good culture, a learning culture? And just another industry, a CPG customer that I work with through this, they were able to identify stockouts and eliminate stockouts by 20%. That is millions of dollars in potentially missed sales from a stockout. So I think this power shift along with the tech modernization and democratized insights is where the true value is.

Rebecca Knight

>> So is it changing the way organizations recruit? I mean, what you're talking about is this learning culture, this growth mindset, this having to learn new technologies and as you said, the people who are doing well continue to do well because that's what they're doing already and the laggards fall by the wayside. But how would you describe how it's changing the way organizations, the kind of skills that they're looking for and the kind of traits and attributes they want in employees?

Cindi Howson

>> Yeah. And so I'll speak industry level and then Mark, you can share from your team, the talent wars have never been fiercer. Everyone wants to hire an LLM expert today. But you know what? LLM experts that existed were in the handful 18 months ago. So what you need to recruit for is curiosity, and who is the fastest learner that is both from the core data job as well as the leaders. The leaders that can combine business, what's their imagination? How can I apply generative AI to this business problem? And they understand the tech stack,. Those are the ones that are much further ahead in this wave of disruption than those that stick their head in the sand and wake me up when this wave is over.

Mark D'Costa

>> What I'm also seeing though, is that the power of some of the generative AI solutions that we have, like within our company there's a AI hub team that's really focused on taking that hill as it relates to AI within the firm and the strategy and they have an app that they've wrapped around these tools called Eliza. And what I've seen with that is that it really brings the access barrier way down. So in a number of instances, you don't necessarily need lots of the specialists and the LLM modelers, what you need are people who are creative, understand their business flow, and have been asking for tools and methods in order to solve problems. So what I'm really seeing is that there's a lot of energy within the firm because there are teams that know their businesses, know their clients really well, and now they suddenly have this additional, let's call it AI assist copilot in order to help them bring those ideas to fruition.

Cindi Howson

>> Eliza, that's an interesting name. Do you know-

Mark D'Costa

>> It is.

>> I wonder if it's the same? Did you know there was a... The first chatbot I believe ever was called Eliza.

Mark D'Costa

>> There's a story around Eliza. I'm not going to try to say it because I'll get it wrong, but think it has some tie ins with our firm's history and some other bits.

Cindi Howson

>> So yeah, they called it a chatter bot back then, and they were amazed like we are today at GenAI, and then it sort of died.

Mark D'Costa

>> Eliza's pretty smart now and she's getting smarter by the day exponentially.

Rebecca Knight

>> Well, I mean, think about Pygmalion, Eliza Doolittle, she learns.

Cindi Howson

>> Okay, now you're taking back to my starring role in junior high.

>> Okay, I want to ask more about disruption. We had a pharma on yesterday, and this individual predicted that there'll be a 30 to 40% disruption in his data pipeline in the next three or four years. Zhamak Dehghani has done a very good job of describing the hyper specialized roles within the data pipeline and the pain of actually getting to that stock out example, you just need so many experts. How do you see Gen AI changing all the heavy lifting that we need to do?

Cindi Howson

>> Yeah. And so two parts, so Zhamak Dehghani, love her work, have had her on The Data Chief podcast as well, but I think more than the tech, what she got people to think differently about that BNY also applies is that domain ownership and agility, that you cannot just have one centralized data warehouse, slow monolithic, all the pipelines going there. So I do think gen AI will help with things like data observability, but also crowdsourcing the business definitions, do you call something bookings or revenues? So crowdsourcing the business definitions both for the pipelines and also for the dashboards, the insights that you're creating, improving data quality as well, looking for those outliers without it having to be so rule-based, those those are the opportunities I see.

Mark D'Costa

>> I also see it. I've seen some really interesting things where it will also lower the barrier to being able to work with data. So I'm a SQL guy and I can do some pretty interesting things with SQL. What I'm also seeing is that with the advent of LLMs and additional tooling, that starts to go away. And so you can start to then have people who haven't had to spend the time learning that technical skill, suddenly they're able to still work with the data, get to insights that they just couldn't otherwise get to before having lean on someone else. So that efficiency is going to be really strong from an outcome perspective. And then within the data pipelines, I do see the ability for much more AI assist in terms of classification labeling, think information classification. Think about 20,000 tables with X number of fields, the surface area is huge. So you need to have some form of human assist there in order to get it right and to scale that out, because otherwise, previously you could only govern a small amount of highly governed regulatory data, now you really want to be able to unleash that data for all sorts of outcomes. You need to have something that's going to help you scale.

Mark D'Costa

>> And I'm not a SQL guy and I don't want to be, but I know my data better than the technical people know the data. And so to the point you were making earlier, Cindi, about the organizational aspects of Zhamak's work, to be able to truly democratize the data you need that human input of those who actually are in the know. And at the same time, it's got to be governed. It can't be the wild west.

Mark D'Costa

>> It does, it does.

Rebecca Knight

>> It's got to be the curious creative liberal arts graduates. Finally, Mark, the year's half over, but I'd be interested to hear your aspirational goals for the rest of 2024 and even into 2025.

Mark D'Costa

>> So look, the key goal for us has been trying to stand up this governed, tenanted Snowflake platform that kind of unleashes creativity within the firm. Also, now adding to that, how do we make a service like ThoughtSpot fully tenanted as well? The data gravity starting to shift dramatically from on-prem to cloud. So as we get to that data gravity shifting, one of the aspirational goals is to start to look at, how do we simplify our lives? How do we start to look at ThoughtSpot's SaaS offering? How do we then... That in itself is going to create more efficiencies and unlock even more features and functionalities that we wouldn't otherwise be able to go after.

Rebecca Knight

>> Excellent. Well, Mark and Cindi both, thank you so much for coming on theCUBE, a really great conversation.

Mark D'Costa

>> Thank you.

>> Thanks guys.

Cindi Howson

>> Thank you for having us.

Rebecca Knight

>> I'm Rebecca Knight, stay tuned for more of theCUBE's live coverage of the Snowflake Data Cloud Summit. You're watching theCUBE, the leader in enterprise tech news and analysis.