CLIMATE. MONEY. WORK. PODCAST | EPISODE 2.1
Data-driven Tech Solutions for Climate Risk
Guest: Gopal Erinjippurath, Founder and CTO, Sust Global
Welcome to Season 2 of Climate Money Work! This season, we’ll sit down with board leaders, Chief Sustainability Officers, regulation experts, and more, as we unpack how they’re each thinking about an evolving regulatory landscape, the explosive growth in generative AI, as well as an updated view on the current global landscape of ESG.
In today’s episode, we’re chatting with Gopal Erinjippurath, Founder and CTO at Sust Global, a tech company developing climate analytics and APIs to enable the climate economy using satellites, geospatial data, and deep learning techniques.
As economic damages from extreme weather events now total $100 billion annually in the U.S. alone, it’s important for both companies and investors alike to gain a deeper understanding of climate risk. We go deep on both physical and transition risk and discuss the role of data and analytics in understanding and mitigating these risks. We also discuss how both the use cases and data sets can differ across sectors.
As Gopal sees it, a world is emerging in the next 10 years where every dashboard will incorporate climate analytics, which is why Sust Global is laser focused on enabling their users to make the best meaningful decision using the most relevant data sets.
Follow: Spotify / Apple Podcasts / Google Podcasts / YouTube / More
Transcript
Keesa Schreane: Extreme weather events continue to take a devastating toll on our global society. Real life catastrophes are becoming all too common from a social perspective, from a human perspective. From a corporate angle, as companies think through their physical risk and transition risk strategies, costs associated with damage from hurricanes, floods, and the like continue to escalate.
According to whitehouse.gov, the economic damages from storms, floods, droughts, and wildfires have risen to over 100 billion per year in the United States. Now, here to discuss how data may be used to progress social, policy, and corporate actions around physical climate risk is Gopal Erinjippurath, Co-Founder of Sust Global, developing climate analytics and APIs to enable the climate economy using satellites, geospatial data, and deep learning techniques.
Gopal, thanks for joining.
Gopal Erinjippurath: Keesa, it's my pleasure to be here.
Keesa: So first of all, let's just start off with an explanation. What is physical risk data? What is transition risk? And if we could go from there to talk about the customers, the use cases, who uses the data for what?
Gopal: Yeah, when it comes to physical and transition risk, many of the times they are bundled together.
But a simple mental model to think about and distill those two concepts are related to the climate, and us as a business or us as a society. Think of physical climate risk as what the climate is doing to us, and think of transition risk as what we are doing to the climate.
So when you think about transition risk, you're often trying to understand in terms of either emissions, or in terms of your carbon footprint or profile. What impacts does that have on the climate and how a greening economy will have impacts on the different pathways into how a business operates in the future?
That's kind of largely what people are connecting into transition risk. It connects into legal, it connects into policy, and if you were to transition to switch to physical risk- in physical risk, you're dealing with how the holdings of a business get impacted by the changing climate. So how does the impact of the extreme climate events that you described earlier, wildfires, floods, cyclones, and then chronic hazards and peril, like water stress, heat stress, sea level rise? How do they have an impact on your holdings, your businesses holdings, and your communities? Tangible physical footprint is kind of what is accounted for in physical risk.
Keesa: So let's talk through the data, talk through who users are, who's your customer, and I assume that's going to be plural customers, and how are each of those segments using the data? So if we had a chart in front of us and we have, you know, customer on the left, and then use case on the right. Take us through each of those scenarios. Pairing up customer with use case for each customer type.
Gopal: Let me start with some prototypical users that have been growing over the last several years in terms of their use of climate data. So I'm going to talk more from the vantage of physical climate risk, but I think that equally represents what you'll see on the transition risk side.
So in the primary users that we've seen grow over the last 10 years is that of ESG or sustainability reporting teams, so teams that are looking at enabling large organizations and across public equities, as well as private equities, to understand what the impacts of climate risk are to their business and communicate that downstream to their stakeholders.
So towards the end, the primary use case for those groups of users is, here's my tangible physical footprint. I wanna be able to understand across many different emergent disclosures, be it the TCFD, in most of your emergent standards in Korea and India, or the SFDR or EU taxonomy in Europe, or be it the SCC guidance, which is now gonna get legalized into a framework for companies above a certain size, they need to go through certain processes in terms of documenting their climate related risk. Just like they have to document other aspects of their business's accounting. So while those come to fruition, the need to report in a consistent way, which is data driven, is getting even more significant and being data-driven as it pertains to their assets, and their locations, and their asset types. So that's one use case that we actively support, and I would call reporting as more like the essential exercise for us to be accountable, as businesses accountable, as a society, into what being transparent about climate related risk.
However, the larger overarching need is for risk management because you can report, but that doesn't necessarily mean you're mitigating or creating adaptation measures. So risk management is the emerging category that we are seeing increasing use of climate related risk data, and by that I basically mean chief sustainability officers, chief risk officers, who are now seeing increasing amounts of risk and thereby losses from the historic incidences of acute and chronic climate-related peril.
For those teams, often it is the enablement of data-driven techniques and analytics that help them understand what the risk profiles look like from climate to their businesses, from climate to their supply chains, from climate to the distribution centers. So those are kind of the more detailed use cases that we enable.
I would say when it comes to risk management, the other spinoff from that is understanding risk mitigation measures. So the use of certain mitigation measures can enable climate-related risk profiles to decrease in certain places, and those mitigation measures could be replicated across the supply chain.
It could be certain adaptation measures that protect against acute and chronic peril, and in some cases, it could be strategic planning that helps plan for the future based on active data that is projecting climate-related scenarios across different kinds of warming environments on an annual, decadal and multi-decadal time horizon. So I would say those are three camps
Keesa: I want you to continue with those three or all of those internal use cases. So right now we're not talking about institutional investors, we're not talking about others, we're talking about how internal CROs or risk officers are using the data. Is that accurate?
Gopal: Right, I would say there is a world emerging where there is a need for consistency in terms of the data that is used internally and the data that is used externally when it comes to asset owners and asset managers who are looking to create portfolios of assets, which have a tangible physical footprint. The internal stakeholder alignment enables them to communicate in a data-driven way to owners and the investor audiences, but many of the times the investors are doing their fact checks and doing their own analysis. Ideally, they were driven off the same data and analytics so that both sides see the same thing. The communication is consistent. So I would say those are, I didn't mention them explicitly as like a user, but the same analytics provided are sold into asset management and asset owners, both for private credit as well as public lending and private lending audiences.
Keesa: So the three use cases that you really pointed out were for ESG reporting teams, and thinking about teams that make sustainability reports, risk management teams, that's an emerging category in terms of climate risk, CROs, and then risk mitigation. So how to protect against these perils. And you're saying that institutional investors also use them, but they create their own models around that.
So from there, let's talk about data management. There is so much data a lot of it might be classified as noise, but there is some great data too. How do these stakeholders go about taking in data, validating that data, auditing the data and then figuring out from their validation and auditing it, what next steps are, how can they take the data and cleanse it?
Gopal: Yeah, it's a non-trivial problem. The main challenge we see most data teams grapple with is the multi modality of the data sets. So when they are looking at a group of assets, or a group of businesses, or a handful of portfolios, or looking at the whole set of listed equities for a specific sector, we're dealing with a lot of different data from multiple different data providers.
One of the bigger challenges is just managing the versioning and tracking of that data. So when it comes to data on climate related risk, The most critical component to remember is the fact that climate data is natively geospatial.
So it's geospatial and it's time varying. There needs to be a way to represent that data with spatial and temporal components, or spatial and temporal dimensions to it, and towards the end, I think when it comes to validating and auditing large data sets, the critical components are abilities to version abilities for representing and indexing the data in a spatial temporal way, that there are techniques and technologies like spatial temporal catalogs as a standard way to expose collections of spatial and temporal data.
Then I think the third bit is probably the most important. A way towards joining that data with the company level information, the company level metadata. When we think about large providers of intelligence into spaces where equities are analyzed, or businesses are analyzed, they are often already in that realm where there's an analytic warehouse of data, but the challenges, they're often in diverse parts of the data lake.
Being able to join them using unified identifiers makes a huge difference in terms of the convenience in doing the analysis in an accurate and a holistic way.
Keesa: So if you're dealing with, data managers who are new to physical climate data specifically, maybe they work with other types of data outside of the ESG set, what are the top three things should they focus on when they're looking at integrating this kind of specialized data into the set that they may have?
Gopal: To be specific on physical climate risk, I think the most important thing is connecting between observational data sets, and projected data sets, and being able to run their own validation numbers and validation tests. Oftentimes, the challenges with working with climate data is climate long-term theory.
And what I mean by that is, when you look at just a specific window of time, be it a week or like a month, you're in the domain of weather. Short-term weather and long to medium term weather forecasting. When you're looking at climate, you need to look at annual, decadal, multi-decadal time horizons.
So being able to access data, which is derived from ground measurements, or derived from satellite-derived data sets, and correlating them with historic projections from models derived from general climate models, or general circulation models, will be a way towards building trust within these audiences as they think about projections into the future.
I would say that's the first thing that a team looks into, which is the level of validation that has gone into the production of this data set. Secondly, the ability to keep that data fresh based on the most recent acute and chronic peril. Data is the second most important component, and then the third is that those two components together enable the audit ability into the actuals, so being able to create a synthetic fusion data product, which brings the best of frontier physical climate risk modeling and near real-time satellite derived and ground-sensor derived observational data sets is kind of what we've seen most teams care about in this space.
The second group of challenges that they are often working with and asking about is how that can be connected into loss data sets, asset-level data sets, and potentially graph data sets. I would say these are the big new opportunities for new data sets in the future, which is climate data sets exist, and they're often at the land parcel level with geospatial context, but the business has its own context, which is, you know, these events that created these kind of losses because my business operates in a certain way.
So that socioeconomic data and commercial data is one to join with these climate-related data sets. The second bit is when it comes to transition risk, oftentimes it is asset level data on emissions, and their carbon footprint, and their ownership.
That's when you're going to scope one, to scope two, to scope three level emissions and methodologies for offsetting insetting, and how they connect in a graph layer, which is how do you maintain the dependencies between different entities in a specific business, and how you look in the connected graph between businesses when it comes to upstream and downstream aspects of a supply chain.
So those are all the adjacent data sets that need to come together for a holistic analysis on physical and transitions.
Keesa: So let's talk about working with your clients to first of all, uncover what their needs are. You talked earlier about the three use cases. Could you talk about how they can get to a conclusion around what their needs really are, and combine that with the data that they already have, to then work with you and to work with others to build an actual product?
Take us step-by-step. What does that look like from validating data, to understanding what we need, to building an actual product or solution?
Gopal: Right, and I would say to get to that, we can transition from the simple to the complex, and I wouldn't say simple is simple, but it's basically saying there's a better standardized understanding.
So when it comes to reporting teams doing reporting, be it for disclosures or be it for due diligence. Oftentimes, there is a process. It's basically well defined from an analyst standpoint, saying this is the kind of data set we need, and these are the dimensions we need within that data set. So there is a specific set of analysis that they're trying to do.
The customer discovery there is often a function of do we have the data modalities to satisfy the needs of a reporting related work stream? That's where I feel the emerging standards from TCFD and the EU taxonomy are helping and bringing together consistency across different teams and their outputs when it comes to reporting. For us, for risk management, and mitigation measures or sustainability measures, oftentimes, those are a little bit more bespoke, so it comes down very much to the needs of the teams looking into this.
If we were to talk specifically about use cases, the risk management for a group of hydroelectric plants that are evaluating their throughput in environments where water is warming and water levels are varying, is very different from a team that is looking at a collection of electric utility assets, which are related to production and distribution of electricity in certain geographies.
Keesa: To dig in on that, what I'm hearing you say is that use cases may differ according to sector. We talked about the different customer types, whether it be internal, whether it be the chief risk officer, et cetera, but now we're talking about sectors. So continue, please. I just wanted to call that out, continue to let us know about the things that different sectors are looking at in terms of data in terms of risk, and what they're trying to solve for?
Gopal: Right, so there are a handful. If you just go with the GICS taxonomy for sectors, from MSCI, S&P Global, there are three kinds of sectors where it's kind of obvious to anyone working in the space that there are physical climate risk impacts. That's real estate, that's energy, that's utilities. So if you just drill down to utilities, you're seeing this clear demarcation. What I'm trying to say is, they all need some form of risk management, and they all need some kind of risk analytics from climate, but they all look at it slightly differently. So if you're a hydroelectric plant, you're looking at water-related impacts when it comes to production. If you're looking at solar, you're looking at a different set of parameters as inputs to your risk management. If you're looking at electric distribution and distribution lines, you're caring more about wildfires, extreme heat, which could disrupt the distribution lines, and which could also enable more fueling in terms of extreme fire, which would lead to a liability risk.
So these are all the things teams are beginning to look at, but the ability to have a layered collection of climate related indicators as part of your product line that you can serve to risk management teams and cater to a broader set of audiences, help you hit the whole sector at once. So that's kind of the direction that we're seeing companies take, which is to be sectoral, look at very specific use cases, create that analysis because we would have to repeat that over the course of time.
Keesa: Let's talk about these sorts of use cases, and products, and the commercial value that they may have. So I'm sure you've worked with betas before and some betas did better than others. So tell us about what the issues were, what the problems were, when there were use cases that were unsuccessful in terms of revenue, whether it be revenue projection, or what actually came in, versus the most successful use cases.
Who were the stakeholders? Were they the correct stakeholders? How did they do things differently to see a better revenue outcome when making these solutions and products?
Gopal: So there’s a class of capabilities, especially when it comes to betas and POC’s, which are related to the audience. So in one camp that I didn't talk about, I talked about risk reporting. I did talk about risk management. There's also teams building new products that enable using climate data that enable these teams. So the distributors of insights and analytics into teams doing risk reporting and management, and oftentimes manifest these platforms.
I would say the engagements that have always been challenging have been when you're looking at dimensions of scale and internal knowledge and integration. So I would say let's just distill those three. So when it comes to scale, if you're looking at just one property, oftentimes you're in a very good spot, and you have your one property or one asset that you're trying to do a very drilled-down analysis of.
Oftentimes, you can hire your own expert staff to do it because there's more at stake. There's more at stake, a heavy amount of risk conservation in just one asset. So then, for the flood-related work and damage, you go to a hydrologist for wildfire-related damages, you talk to a wildfire expert, but as you can see, that doesn't scale very well when you're looking at hundreds of assets, or thousands of assets, and that's what oftentimes, large asset managers deal with which is their holdings are not in non property, their holdings are distributed over many companies that have thousands of assets. So that's one dimension.
When you have a very small number of assets, you're better off going with some expert teams if you have the budget for it. Otherwise, you can use an offering which is more bespoke and more focused on software as a service model. The second dimension is that of internal expertise. So many of the times, teams don't have analysts on staff , or analysts available for translating the data around climate into meaningful reports.
So that's kind of a meaningful analysis that talks to their stakeholders. That's an area there, I would say, is not a challenge, but it's been an area where we've had to do very close collaboration with their team saying, here's the data and analytics, here’s how you transform that analytics into insights into my audience, and oftentimes, only the partner team knows and understands that media. Which is why I think advisory businesses have a very critical role to play when it comes to the enablement of climate related data and climate related analytics into reporting disclosures, as well as risk management strategies, as well as new products that are getting served into spaces where climate is still new. There is awareness around it, but there's not a deep understanding of what climate can do for businesses.
So that's the second bit, and then the third dimension is that about just learning. So many of the times people confuse climate with weather and just enabling the learning and knowledge sharing in a growing ecosystem of users. It's an area we've had to invest in significantly, which is true with most new technologies.
Keesa: Can we dive deeper into the advisories piece? I thought that was quite fascinating. So are you saying that there is a level of knowledge that's missing in many of the advisories? Are you saying that advisories might be a good place to start if you're a company who's looking for some domain insight?
Gopal: Right, I do feel like the advisory businesses provide the bridge between software as a service products and platforms that provide a range of different insights, but they can enable translating that into the core needs of a business. So many of the times for like a tech-based venture, getting into the specific applications is doable for a handful of businesses, but then there comes a time when the advisory business is the one that can help bring in that domain knowledge and make the process more smooth and enabled.
Oftentimes that's what teams, who are looking at this from scaled analysis care about, which is, say, I have all this data, but I also want to translate that in a meaningful signal for a decision on a specific location.
So you're going from a tangible set of equities. To scale the analysis, to assessment, and risk management at the equity and at the asset level. So in the last mile of that, they're all often humans involved. It's not just a decision based on a set of numbers that are served from an analytics platform.
That's an area where we do a fair bit of collaboration with teams that are very well-informed on the domain.
Keesa: When we're talking about data strategy, we hear that a lot these days, “you gotta have the right data strategy”. What are the top areas that you would recommend if you're working with a firm in terms of data strategy that they need to consider? What does data strategy look like in this situation with these use cases?
Gopal: It's a great question. I feel like the primary needs when it comes to strategy is alluding to some of the things I mentioned earlier. Being able to bring together different data assets and provide a common framework for thorough analysis. In that sense, I would say that dimensions to the data strategy often surround freshness. Which is, how can you bring the most recent data in? How do you enable reproducibility? Which is if you've done an analysis, being able to do it again, like the data set from a particular point in time and then redo the analysis when new data is coming in.
I would say, these are a little bit more tactical because many of the times when people say data strategy, we automatically think about data lakes and warehouses, but I'm trying to talk about what really matters regardless of where you store the data. So its reproducibility is enabling freshness in your analysis and enabling a balance between on-flight compute operations and preprocessed data because they tie into two different modalities in which climate data can be served. Which is, you can bake the data into a set of assets or a set of locations. So climate data at county level, but oftentimes that doesn't enable a bespoke analysis related to a group of properties in the same county, which can have very different kinds of risk profiles. So then you can do on-the-fly computing operations.
So where do you draw the line between preprocessed data and on-the-fly computing? That's part of your strategy in terms of working with climate data because you understand what kind of analysis that you want to do. Lastly, I would say data representations in terms of speed and access over large collections of data, which is what's the best form of representation.
All these are things for data teams to consider when they think about data that they bring from external providers into their analytics-driven warehouses.
Keesa: Speaking of working with external providers, let's talk about pricing. If you're working with, you mentioned advisories working with consultancies, you know, that's one of the first things that a client asks, what should my pricing strategy look like? What kind of budget do I need to have? We can talk broadly here, but let's talk about pricing and how pricing might differ according to use case or according to size. If you give us some ballparks and the differences there, that would be great.
Gopal: Yeah, I would say there are like three dominant models and there is a bit of overlap between them, but the three themes I see consistently recur is when people see a software as a service product, or an analytics product, they think about subscription, which is can I subscribe to this product and have all you can eat kind of consumption? So that's one modality.
Then the pricing is over what's the universe of assets or regions or entities that you want to map your coverage over, and then it comes down to the term and it comes down to the distribution. So the pricing is often a function of that. The other model we have seen is around consumption and that kind of connects more into consultancies who are still mapping out the spaces into which data can be so, and for them, it's often a function of usage over time. So oftentimes, that's where the pay-as-you-go kind of model often works, which is you pay every time you use the product and the pricing for that will be different.
So both of them tie into two patterns of data usage, and two different patterns of pricing. I would say that's kind of what we're seeing across the board. So if it is a platform level access to say, you know, a big part of our business is serving data through APIs, and that's one of the things we're uniquely good at, which is standing up the capability and serving this data from a geospatial or an asset level standpoint, through entity identifiers into larger platforms through an API.
We can serve both those patterns, which is the subscription pattern over a universe of locations or a universe of regions and the consumption pattern, which is over a universe of assets. On nature based, real estate assets, or nature based assets, which can be on demand and consumption driven.
Keesa: Let's talk more about those nature based assets as well as biodiversity data because you hear over and over again that it's just hard to come by. So talk to us about what you provide and again, use base centric. What is the use case? When would nature related or biodiversity data be needed? How are customers using it?
Gopal: I would probably just distill down to the nature of finance aspects. So across the financial ecosystem, there is an increasing interest to understand risks associated with nature-driven assets. So one good example of that, the carbon markets, are nature based carbon solutions often manifesting in the form of reduced degradation of forest or plantation of new forest, which is afforestation or plantation of new kinds of coastal resiliency and coastal results which are often mangroves. So those are all natural assets that need to be developed, protected, and maintained over a long-time horizon for carbon sequestration to happen to naturally persist.
So two users in that space in the community of nature finance are nature based carbon project developers and nature based carbon project investors. The biggest challenge facing those communities is the trust component. So developers are seeking investments for developing their projects.
Investors are seeing an ocean of projects with debatable quality or debatable kind of investment opportunity risk. So we wanna understand the risk a little bit more and build trust in the investments. The biggest impacts that we are seeing to nature-based, financial, instruments serving into the carbon markets is the presence of wildfires, and a drought in forestry projects, and the presence of sea level rice tsunamis. As well as cyclone related risk to mangrove projects.
So what do we do then? We enable the assessment of these projects from a climate risk standpoint for their durability, and their long term permanence, for the credits issue and thereby the carbon offsetting possibility for these projects over the long term.
So we're seeing nature finance, the increasing alt investment strategy across the asset management universe, for those asset managers, they don't need to leave nature finance experts when it comes to forest to a blue carbon, but they want to understand the risks in terms of your participation.
So how do we help them distill the concepts of quality and trust?
Keesa: So with asset managers, we're looking at those risks that you just named. They want to know basically what they're getting into. They want to understand what the coastal resilience, coastal reserves, forestation, that sort of thing. I'm also thinking about insurers, right? So we have the asset managers, you have the developers who also want to know this background so they can know what they're building. Talk about insurers–do they have the same use case or do insurers care about something a bit different?
Gopal: Yes, the insurers are looking at a combination of things, and I would say in the nature based ecosystem, the insurance landscape is maturing. It's still very different from real assets, and there are emerging new players, but they're often looking at risk from a multitude of factors. They're looking at many of the times you have a forest carbon project. That was started in Ecuador, and the most dominant risk is not just climate, but also political risk.
So how do you look at risk holistically? Like they're often looking at the problem of how do you look at risk holistically across the political dimension, across the execution, and the execution risk dimension, and also importantly, from the climate dimension, whereas with our capabilities and our ability to use a combination of remote sensing and climate modeling powered by AI techniques.
We are very focused on the climate risk dimension. So we are enabling multiple of those providers with the ability to look at nature based assets and nature finance from a climate risk standpoint, while they're focused on the other dimensions of risk.
Keesa: Now, let's talk a bit about just your work and your experience with Sust Global. You've talked broadly, but I'm wondering, what are some of the things that as a customer, why would they want to work with Sust Global in terms of your ability to scale with them, and in terms of your ability to be flexible?
Let's talk about some competitive differentiators. What do you see as key better differentiators, and how would you say that Sust Global really can meet some of those?
Gopal: Yeah, I would say we're like a four year old business that has been very focused on bringing together the state of the art data and analytics into the space where climate related risks are just the most important things on the top of the minds of our customers.
When they are often looking at this, they're comparing it against catastrophe modeling outcomes which have created losses for them. They're looking at a range of bespoke solutions that don't scale across geographies, or across time windows, and they're looking at solutions which are providing very high level scoring without the ability to drill down and distill into more detailed metrics, and that's where we're bringing together climate related data when derived from remote sensing when it comes to historic impacts and historic actuals, we are bringing together climate related models, modeled results, which are informed and validated against these observations. We are packaging that in a way that is very much ready for analysis and integration.
So that's kind of the main differentiation we have seen across the marketplace, which is, you know, everyone standing up a dashboard with high level summary scores. We believe that there's a world emerging in the next 10 years where every dashboard is going to be climate informed into the audiences and spaces we serve, and all of them will need climate data and analytics served up to them because they're not natively climate teams or climate data teams. So how do we access them, how do we enable them to make the best meaningful decisions using the most relevant data sets?
Keesa: That's interesting you talk about the dashboards of the future, everything will have climate data, and climate data will be infused into all these dashboards.
Do you see the future as one where customers are primarily focused on feeds, so they may have their own platform that they've built out, and they'll get feeds from these companies, or do you see a future where customers would be relying on various different, you know, software, very certain platforms where they would have vendors that would need to have the ability to feed into those specific platforms?
So I'm asking, is the future really oriented to feeds that come directly to a customer-created solution or is the future oriented toward platforms that are built, and customers really gravitating to those maybe five or six vendors that build these platforms that enable the solution?
Gopal: I think I see a blend of those two worlds and let me explain what I mean by that, there are certain platforms where the feed is going to be important. It comes down to how you're distributing this sort of data and these sort of analytics. So there are existing analytics platforms used across the financial services ecosystem, and to them, this sort of data is always served as a feed, but there are certain times when there are teams trying to build their own internal dashboards because of the bespoke data they're working with, or the very specific workflows they have to work with. Now they have to build out for decision and business intelligence internally and to their stakeholders.
So in their cases, there would be the opportunity to go directly. Not to the platform, but directly into the decision making engines. So you're seeing a mix of both. I would say the platforms are moving pretty fast in terms of adopting and integrating climate related data, but there's always a trade off between general data sets and bespoke data sets, and platforms will always tend more towards the generalized data sets.
So that's where there's still the opportunity to go direct and go to the bespoke analysis, which is very specific to our businesses.
Keesa: So in terms of bespoke analysis, and this kind of ties back to the pricing question too, do you see many firms looking to get smaller subsets of the data? So if we're looking at all physical climate risk, maybe some firms in certain regions or who have a focus in a certain region may just be looking for drought as opposed to all of the other weather conditions.
Gopal: Yeah, there is a regional element and that's a big part of, you know, when we started out looking at this problem, we natively realized that for doing comprehensive analysis around climate, you need to look at a range of different indicators and you need to look at global scale. What that means is, we produce the data at a global scale and provide analytics at global scale, so from a range of different indicators, which are climate-derived.
Then it sells a broader category of use cases because someone's looking in just going back to the nature finance use cases, the blue carbon projects are often not worried about wildfire, and the forest carbon projects are very rarely worried about things like flooding or tsunamis because they're not in coastal areas.
So there in comes the region-specific and type, or asset-specific types of risk that they are interested in. So when you have a global data set with many indicators, you could then serve those broader categories of use cases.
Keesa: So let's talk about the tools that you're using. We're talking a lot about A. I. These days. Obviously, we talked about geospatial data. You talk about satellite data. What are some of the new ways that you plan to use these tools, or what are your customers asking for that's forcing you to look at tools a little differently and how you use them a little differently?
Gopal: Yeah, it's an evolving landscape. So I would say it comes down to the data product. So I would say I'm going to start general and go down specific to what we're doing. When it comes to any data product, which is AI-powered or a data transformation, or transform data product, you're looking at freshness, which is often connected to frequency. You're looking at coverage, which is often connected to spatial in the geospatial context, or a group of the index or instrument you're trying to map towards. You're looking at layering, which is the collection of indicators. You're looking at tiering, which is going from summary level, to asset level, to asset at time series level, and then composability, which is like saying, okay, I have the data, but can I join it with other data streams that provide very meaningful financial insights?
So that's where we've been very focused on, which uses AI to enable freshness into the data product. Use AI to provide coverage and by freshness, we mean, how do I enable the most recent climate related observations to model the 21st century climate conditions, that enabled 21st century modeled climate related hazard projections. So that's what I mean by freshness, which is a type of frequency.
The coverage space is very interesting, most global climate models are modeled at a very coarse spatial resolution, whereas most of the use cases that we have discussed are very specific to assets, be it nature driven, or be it tangible real estate, or industrial assets. So they're having spatial coverage at high resolution for the point. Layering connection to all the different hazards you're trying to map into and many of the times, they're all derived by the same set of fundamental climate variables, which is, temperature, precipitation, soil moisture, and a range of others.
So how do you bundle them together and use AI to create indices, which are tied to very specific use cases. So that's the layering component. When it comes to tiering, how do you bring together many different indicators to say, here's the risk related to wildfire to a property? Here's the risk related to flooding the property?
So that's kind of the tiering component. Then I would say the composability is more on the interfaces and the API side, so a little less driven by AI, but AI has a big role to play across the state of products, and we've been very proactive in terms of bringing our AI driven smarts to the community, sharing our methodologies in an open way, and enabling a deeper understanding of the analysis we do, bringing together remote sensing data and climate modeling data into these spaces, be it real estate or nature finance.
Basically the dimensions I look at from a data product standpoint, freshness, coverage, layering, tiering, and composability.
Keesa: So just in wrapping this up, I always like to ask, tell us something we did not know. I want to ask you specifically, tell us something we did not know about the key skill set that's going to be needed in the future in terms of aggregating this type of data, and make sense of it for folks who are in the industry, want to progress in the industry, want to maybe being inspired by you have a startup in this industry.
What are the key skill sets that are needed to be able to analyze the data, make sense of it, and even communicate it to clients? Or if you have a top three or top four, these are the top things that are needed to analyze data and communicate it to clients and build successful products in this space.
Gopal: Yeah, I would say the key takeaway is natively working on climate data and analytics is a multidisciplinary endeavor. You need to bring together skills from the climate sciences skills in the data and analytics realm, and skills derived from domain-driven, like the specific domains you're serving into. Oftentimes, all that doesn't exist in one individual. So many of the times people hesitate about getting into the climate because they think, “Oh my God, I don't understand it well enough”, but the reality is there's a new space, everyone's learning together, and everyone's coming up to speed at the same time. So you have an opportunity to bring your unique skill sets in whatever aspects of the data ecosystem you work in, into applying it for climate, rather than it being you're a climate expert trying to learn all these other skills.
Keesa: I mean, wow, how you wrap that up so fantastically. There's a regulatory piece that you need to be aware of, there's the ability to be able to communicate with customers, understanding their needs, and being able to translate that into a product. There is the climate science piece, but then there’s as you mentioned, multidimensional being able to connect dots, whether it's geopolitical, whether it's climate or whether it's engaging internally with stakeholders. Fantastic, fascinating conversation, Gopal. You and I could do this all day long because this is just what we do. Thank you for joining us.
Gopal: Thank you, Keesa, I appreciate it, and you know, very glad about the great work you're doing, with your book, Gambling on Green. I do feel like many of the things we talked about are kind of the precursor to enabling intelligence into Gambling on Green, which is how do we make the right choices as businesses? How do we make the right choices as a society towards enabling the most sustainable future that is through data-driven decision making?
Keesa: That means so much to me. Thank you so much, my friend, Gopal. Thank you for joining and thank you for that.