Sponsored by: ThoughtSpot | How Chevron Fuels Cloud Data Modernization

Summary

Cloud migration strategy: This video shows how Chevron transitioned central finance and procurement analytics to the cloud using the Databricks Data and AI platform and ThoughtSpot's agentic analytics.
Legacy system modernization: Learn dos, don'ts and best practices for replacing outdated legacy BI tools with real-time, AI-powered insights on the Databricks Data and AI platform.
Business impact through analytics: Discover how enhanced semantic layers with user-driven understanding drove more impactful strategies for customer engagement and business growth at Chevron.

FAQs

What challenges does legacy BI create for organizations?

This video explains how legacy ERP and BI systems create data silos, require specialized technical expertise to use effectively, and lead to spreadsheet-driven processes prone to errors. The Databricks Data and AI platform and modern analytics solve these challenges.

How do I transition from legacy BI to cloud analytics?

This video shares Chevron's migration journey, best practices for consolidating multiple ERP systems and strategies for democratizing data access on the Databricks Data and AI platform so business leaders can self-serve insights.

What is agentic analytics and how does it help?

Agentic analytics uses AI to enable natural-language queries and autonomous insights discovery. This video demonstrates how ThoughtSpot's agentic analytics on the Databricks Data and AI platform makes insights accessible to all users, not just technical experts.

What business outcomes did Chevron achieve?

Learn how Chevron improved data access, eliminated spreadsheet processes, reduced reporting cycles and enabled faster decision-making for finance and procurement teams by modernizing analytics on the Databricks Data and AI platform.

Full transcript

All right. Hey everybody. Uh, welcome to our session. Uh, how Chevron fuels cloud data modernization. Uh, today we're going to be talking about the transition that Chevron's making to a cloud-based future, moving away from our legacy ERP systems and a lowinformation environment into a highinformation environment. Um, and then we're also going to show you some of the capabilities that THSpot has in a live demo. All right, so the speakers today are me.
I am a insight delivery analyst at Chevron in the finance technology services group. Um, and we also have France, uh, SVP of product management at THSpot is going to be giving you a demo after me. So the before times, how everything used to be before we started moving towards a cloud-based intelligent future. Our legacy ERP and BI systems have always been kind of a sore point for our finance team, right? We have multiple ERP systems.
We're not working with one unified system that can give anybody all the data. We have different instances of SAP. We have GDE. You know, we're a global company with different business streams. We have ships. We have oil wells. Obviously, we have all different kinds of processes. And this led to a uh real uh spreading out of the technologies that were used over the years. So we have a lot of different ERPs that don't really talk to each other.
Um in addition to that we have existing legacy BI tools and these you know legacy BI tools can be flexible but they don't have the ability to democratize data. You need somebody who is uh really good at using them to make them effective. They you and you can't really have somebody that is a business leader that's not a technical expert use them. You also ended up with a huge amount of spreadsheets and emails, right? Everything was downloaded out of an ERP system, put in an Excel sheet, emailed to somebody, somebody put some comments on it or some numbers in it, they emailed it to somebody else, and then you've got three versions.
You know, you're supposed to be using version three, but you were actually using version two. And you know, that just goes on all the time. So, this is not a sustainable environment. And it led to a few problems. One is access and expertise. As I mentioned a little bit, you know, information or data is not useful in its raw form. You need to turn data into information from which you can gain insights. And when you have multiple systems, you can't do that easily as a regular user.
You needed to have access to all of the different ERP systems and the way to get the data out of all of them which nobody really had. You also had to have a lot of extens of domain knowledge. You a person who didn't have a deep understanding of what the data meant wouldn't really get anything out of it. They can pull all the data they want, but it's going to be codes they don't know the meaning to, cost centers they don't understand. it's going to be just a lot of uh you know jumbled information that's not going to give them any information or generate any insights.
So you really had to have somebody with all the access and expertise that was required. You also had a problem with response time. If a leader of the business had a question, they would need to send a note to the finance manager who would then delegate this out to a finance analyst. They'd then go in their ERP system, download all the data they needed, put in Excel, maybe made a pivot table, did some calculations back to the finance manager, back to the business manager, and then finally they would have the answer that they were looking for.
You also had a problem with response times in our legacy BI tools. You can't just quickly on the fly bring in new information that's not already pre-built in the dashboard you're using. You need somebody to go in in the power user version and make the changes, add in the new information, hope that it's what you're actually asking for, send it to the analyst who then looks at it and might say, "Oh, this is still not right."
Sends it back to the power user is going to curate the data. So, all of this leads to a low information regime. The business leader starts to realize, oh, if I ask a question, it's going to take four, five, six days for me to get an answer. In some cases, it's going to take a long time. and they might even tell me they don't even have access to the information they need to give me the answer. So the business leader stops asking questions or they ask only the bare minimum and all this leads to the low information regime where the because of the time spent gathering information and how hard it was to get the information to the right person.
People just went without. They would only use the absolutely mission critical information that was that they knew they could get that they had existing processes to reach. um and nobody was able to make any enhanced decision- making because of that. But today we are moving in a better direction. Our finance department now has the ability to use data bricks and thspot together. What this does is it brings all of our data from all of our disperate ERP systems into one data lake.
Right? That data lakeink is virtualized and accessible by data bricks. So now all of the data is available. Anybody who has access to the link can get it. This is not always helpful though for everybody in the finance department. People in finance are generally not SQL experts. They are generally not well-versed in data structures and metadata and how it all works together. So then we use thspot to work together with data works.
With thpspot they can access the data in a natural way they understand. You can easily put in synonyms, whatever comments they like to see. You can easily visualize the data in a way that makes sense to the user and they can access the data using natural language. That's really key for somebody who is not a data expert. So these two together allow us to have a datainformed decision environment. Right now people can quickly and easily get get access to the information they need to give the answers that people are looking for.
So if we think about the regime that we talked about before in the low information regime and flip it on its head, think about it from the perspective today. Right now all of our analysts can access the data in real time. The tools that thought has are incredibly flexible and they allow the analyst to create new calculations on the fly. They allow people who are not analysts at all who don't understand finance data allow them to ask the question in a natural way.
So democratizes the data. This also helps with response times. Leadership can engage directly with the data and ask questions in natural language. So now instead of sending an email to the finance manager, sends an email to the financial analyst who then looks at the data, hopes he has it, sends it back to the finance manager, sends it back to the business manager. That's not happening anymore. Now the business manager can just ask the question directly into the tool.
He can go to ThSpot and type what he wants to know, give him an answer. Now the analyst is not spending all the time generating that answer. They're not spending all the time pulling information out of an ERP, manipulating it in Excel, trying to make sure that they have the right data, trying to help uh update some existing dashboard. So, they can take this free time they have and perform deeper and more impactful analysis.
Now, they can think about, you know, well, what would the business manager need to know that they don't even know they need to know? or they've been afraid to ask because in the low information regime they knew that it would take too hard to get an answer. Now the analysts can do their job much better than they could before. They have more time and better tools to do their job. So with all of this extra time and all these better analyses, we move into a high information regime.
Now the business decisions can be informed by higher quality insights that the analysts can generate using these tools and better decisions will lead to better operational performance. So that it really does have an impact on the ground, the ability to get this information. So that's where we're at today. But now it's time to think about the future a little bit because nothing is standing still, especially in this space as you all know.
That's why you're at the data bricks conference. So the future right now is about agentic AI. In the future, you as even an analyst will not have to work as hard to generate insights. you will not have to hope you're getting insights that are useful or they are correct in terms of they have the right information. Now going forward we can start using things like spotter spots agentic AI system. So what used to be hard what used to take a lot of time and effort to get an answer to can be done using very very natural language.
We have a little gif you can see on the side here. You you can see how you can talk to the agent and just say, "Oh, no. I actually meant this." And it'll instantly correct the query that you were trying to craft. You don't even need to know SQL. It'll craft the SQL for you. It'll pull the data. It'll really make everything uh a lot better for the analyst doing the analysis. They can also use the agents as thought partners.
So, you can ask the agent, okay, what should I be thinking about? And even if it doesn't give you the perfect answer, oh, this is the exact ratio that the business leader needs to know. It might get your your thoughts moving, right? A lot of times when I talk to some kind of LLM, it'll tell me something and I'll be like, well, that's not quite right, but now it makes me think of this. So, it's a thought partner with the analyst.
The these agentic systems will help the analysts come up with new and novel analyses that they didn't have before. We also be able to make these things work autonomously and have them use reasoning ability. And it will be a really important path going forward for all functions but I think especially for finance. It's going to be a core competency in the future. It's it's not going to be in the future you will not be able to be a successful financial analyst unless you are using these tools to their utmost capabilities to understand how you can use them to better inform business decisions.
So now I'll talk a little bit about some of the things that are important to have a successful transition from a legacy uh low information ERP regime into a modern cloud-based one. So I'm not going to spend uh any time talking about the technical aspects of it. Um it's easy enough to do that. Um it's you'll be it's it's pretty it's pretty incredible how easy it really is to do. So, what I'm really going to focus on is more of the organizational change aspect.
How to work with your uh partners at your company to try to get things done. You do want to make sure that you engage with the customers to understand what they want. And this doesn't just mean ask them, you know, what kind of reporting do you want? Do you want a new report for your operating expenses? Do you want this or that? You really need to have them as partners. Try to create something. Make sure that you have the exact data they're looking for.
Work iteratively. Make sure that as something comes out, you say, "Okay, here's our MVP one. Is this what you're looking for?" They'll probably have some feedback. You got to make sure you're working handinhand with the customers to understand exactly what they're looking for and how you can make their lives better. The other thing that's important is to work with your data architects to understand your data environment.
Uh so earlier this year we uh did a new PC with THSpot and a business unit and our data team was tied up you know a lot of competing priorities. So we just decided to go and try to build it ourselves and it's kind of a testament to how uh easy to use tools like data bricks and thspot are that we built the whole thing out and then once the data team was available to work with us they looked at it and we almost had it right but not quite.
The important part is that there's just that, you know, 10 15% of stuff we were, oh, this is not exactly the right table. You should have been using this one. Um, and this is the newer view. This is the newer hierarchy. Whatever the case is. So, as easy as the tools are to use, make sure that you also work with the data department um to make sure that the data is absolutely the best data that you can be using. It's also really important to have a blue sky attitude.
A lot of people get stuck in old ways. They think about what's always been done, how things have been done in the past, they don't really consider the possibilities of doing something in the future. Um, when they don't think about what's never been asked before, it's really important to just say, okay, in a perfect blue sky scenario, what would I want to see? How would I build that? It can really open up a lot of uh opportunities, and you might even find yourself going down paths that you didn't think about in the first place because your other idea led to that one.
So just keep an open attitude towards everything. The other thing is important is to develop an understanding of the new capabilities. So you're going to have a lot of new capabilities that you didn't have before. And to effectively uh get buyin from customers and make the best possible products, you need to understand how these things work. Um know what they can do um and show people what they can do. There would be a lot of folks that don't really understand what is possible in the new regime.
So it's really important that you develop your own understanding so you can effectively communicate that. Now on the other side we have some things not to do. Don't think don't build things speculatively hoping that you'll have a customer. Um one of the first things that Chevron built with thoughtpot they guessed at what a customer might want and they were wrong. So they built it, took it to the customer, and the customer said, "Well, that's not what we need."
It's really important that you have a customer involved that'll make everything easier and better from building it out to getting buy in to generating value. Make sure that you have a customer that you're working with and not just building speculatively. Also, don't try to recreate the current state. There are some people are going to say, "Oh, well, we've always had this report built out this way. We want to see it like this."
Challenge that. Do you really need it like that? Is it absolutely crucial for you to have it in that exact format? Are you okay with the same information maybe laid out a little bit differently? You know, there's a lot of folks that I work with who have these Excel spreadsheets with like multiple merge and centered columns and all kinds of formatting and they don't want to see anything any other way. Uh, and that can lead to a lot of roadblocks in a lot of different areas.
So, you know, challenge the assumptions about what's always been done. I you know m make sure that you're working towards what's needed not what is wanted because it's the easy path that's always been on. Also don't make assumptions about what customers want or what is possible. This sounds kind of similar to the one where you have to engage with customers to work what they want uh to understand what they want. But the difference here is that you don't want to assume that you that what you're building for them at their request even is what they actually want.
It goes back to that area of engagement with a customer. They say they want this piece of data. Make sure you're working with them to get the right piece of data. They're the business experts usually in that group. They will understand what they're looking for better than you as someone who doesn't work in that business directly. So, make sure that you're working with them constantly to make sure that the information you're getting is right.
And finally, don't hesitate to evangelize and demo. Um, the more people that see the advanced new tools that we have, the more people want to get involved with them. A lot of people have seen generic demos, but when you can build a demo for some group using their own data, it's much more impactful. Um, it really makes a world of difference if you can build out a PC at a customer's request using the customer's data. It really does help the customer see what could be done in this new environment.
And don't hesitate for people that say, "Oh, no. I don't care. I just want to, you know, offtheshelf demo." Show them their data, too. That's fine. But just make sure you get out there because our experience has been that the more people see it, the more people want it. So, just make sure you're you're evangelizing this new high information regime to everybody that you can. All right, that is the end of my remarks. Uh, now we will have uh show us some of the thoughts about capabilities.
Thank you. Thank you. Um, so who already heard about spot in the room little bit? So let me give you like a short introductions and then I want to show you a little bit what you know all the goodness that was speaking about how we can bring you inside uh easily. So Sspot has been around for quite some times. Uh we are leader in a magic quadon from Gartner in uh analytics and BI. Uh also we have lot of great reviews on G2 and peer review.
We have more than 1,000 customers today using us. Now you are going to ask but what who are the customers right? So here we are. So we are cross industries uh big names obviously uh smaller company also we are extremely proud that you know they trust us with uh with their business so you know BI is not new right when I was not even born in France we had like business objects I don't know if everybody remember business object right exactly right so uh it was it was a great BI company sometimes ago.
Actually, some customers are still using business object. Believe it or not, it's great. Uh anyway, so you know, the BI world starts since the beginning from the visualization point of view, right? I want to be able to visualize my data. I want to have a dashboard and this are great and you know, don't get me wrong, they are not going away. They are here to stay. But at the same time, and this is really was the the philosophy of hotspots, how can we build a system like you are on Google, right? when you want something you search and you find what you want.
So that's really what the principle of satspot how we can allow our customers to search data right and obviously with LLM it's even bringing us more technology to go further out so now we are embedded uh LLM technology to provide an agentic experience where you can really like ask in natural language and get answers and for us really this is a combo the winning combo for our customers providing visualizations but also agentic experience all together and of course we also looking at the future with autonomous which is something that we are working on uh next how we are deploying this technology right uh we have six main components so the first one is AI agent everybody should have a personal assistant to answer data questions then the dashboard as I mentioned they are not going away but it's not because they are not going away that we cannot mo modernize them which we have done so we have infused use AI in our dashboard capabilities that now it can provide you insight uh directly on the dashboard you can drill down I will show you that finally all this goodness a lot of companies they actually have customers also and they want to be able to share this insight with their customers so that's why all all our product can be embed in our customer uh application for their customer directly and so you cannot power everything like that without a data layer which are the three other uh dots on the left.
First data management right so we want to make it easy and simple that's why we connect directly on top of your data warehouse data bricks for example so you don't have to move the data you can run all the dashboard you can write the agentic experience without moving any data thanks to a semantic model so the segment model is really like the key of all the puzzle through the segment model you are going to be able to run the right SQL statement to your data bricks but as a user you are going to be able to describe all the tables, all the relationship, all the description of your informations that then are going to be used by the dashboard and the agents.
And finally, we don't forget the data scientist people. We know that there is uh they need to be able to write Python notebook. They need to be able to run SQL queries. They need to be able to prepare a little bit more their data. So that's why we have analytic studio at the top of the pyramid. So thank thanks to this uh agentic analytics platform there is three things that we want to do right as I uh mention it. First we really want to offer a dedicated analyst to everybody.
Everybody should be able to ask their question in natural language. Secondly we want all the insight from the dashboard to be connected to the people. So we don't want necessarily them to go to a dashboard to hotspot. We want to be integrated in Slack. We want to be integrated in steams. We want to be integrated in Google Slide, PowerPoint, uh in Salesforce if you're the Salesforce service now and you know name it, right?
It need to be connected to the people where they are. And finally, as I mentioned, we need to be able to bring this intelligence into everybody applications. So if as a customers you are creating your own applications, you want to be able to embed this analytics for your customers directly because every company at some point is a data company and you need to be able to leverage all the data that you have in your uh environment.
So no more slide. Let's go to a demo. Who is ready for a demo? Good. Thanks. So let's try that. It's the first time that I have subtitles here. I think it's really cool. So for once people can understand what I'm saying. So it's kind of nice. Um anyway let's start. So everything is live. So hopefully everything is working well. Okay, here we go. Okay. So oopsie. So you know you are in a conference obviously you don't have your laptop.
So what do you got? You got your mobile phone here. This is my mobile phone. I have my spot application. Here we go. See live. Okay. Oops. Sorry. So, here we go. So, it's landing on my what we call the watch list. Watch list is like the list of KPI that matters the most for you. So in this example, let's say I'm a VP of customer success, right? I care about churn with churning, what are the support call, support volume and everything like that.
I can see everything here from my mobile phone right away. If I click on any widget, I can see like you know why the numbers is changing and I can see the drivers behind the change here just right at the bottom. So everything is at my fingertips. But I can also have access directly to spotter my personal assistant. So here I can ask him for example and you can see some pre-an question right uh monthly trend for customers.
I click here and this is going to run my queries right away on my my mobile phone. But as I mentioned before right board are not dead. And here it is my lifeboard my favorite liveboard. I can access it here and it's available. And what is really amazing with this application is you don't have to create like specific setup for mobile. You don't have to create layout for your dashboard for mobile. Everything out of the box is working right away.
Now let's arrive to my to the office. So I go back here. We go. Oops. So now I go to my desktop and here is an example of dashboard that you can get in sspot. So it looks like a normal dashboard like you will tell me like yeah you know I have the same in Tableau PowerBI blah blah blah but you know actually not really there is lot of great features that we have added there. So the first things that I love is when I go here I don't want to read all the numbers.
I don't want to understand what's going on here. I just want to click on one button AI highlight. And with this button automatically I'm going to have a summary of it. Summary of my dashboard. Here we go. So prona you need to know that we have unexpected change of support calls. Uh the numbers is uh dropping a lot more than expected. And here are the reason why. And by the way the churn rate is kind of like normal and okay and and so on.
So a personalized summary I can subscribe to it by email also. So every Monday morning I will get all the information uh before I arrive to the office. But also every widget is smart here. So for example you know every time you have like a question on the dashboard like hey this number is going high but actually uh what are the churn reason. So here I don't need to call somebody to change the dashboard. I can very simply click drill down and I can select the CH reason and right here without doing anything special effort I have now all the CH reason for this specific month and I can continue here oh service issue can I do it by uh state to know which state have specific service issue and here we go so looks like if you go to uh Illinois it's not so great so that's really like what we speak about dynamic dashboarding that is really cool.
But also if a number is changing I can click and I can select oh why this number is increasing like here again one click in analysis analyze change I can get a change analysis and I have all the drivers that are changing my numbers here all the primaries and are available and it's not done right I have much more capabilities I can click on spot IQ and bring a full team of data scientists next right to me I can do forecasting I can trend analysis, I can do outliners, everything's embedded inside my dashboard at a at a click at a click away.
Um, and you know, as you can see here, all the visualization are are really cool because we are actually using our own own uh visualization libraries which have more than 40 different chart type and if you don't like ours, you can even create your own. So, it's really really really powerful. And finally, you know, if you don't like to drill down because you're kind of lazy, you know, you still have sporter. So sporter in one click, I'm going to be able to ask question on this widget with natural language.
So again, I don't have to bring somebody else in the team and I can say, okay, you know, um what is the uh churn reason for let's say March 2025? So and right away from there next you know it's same if I will have some analyst next to me I'm going to get the answer right away and I have my answer and still we are using what we call the search token which is a representation of the SQL query because again we are targeting the end users end users don't know SQL like who knows SQL in the room here are you like familiar see like you know very few people so we want to speak to everybody here so that's The search token is really like an expression of it that allow you to validate what sporter is telling you and say yeah okay this is looks correct or no don't look correct and then you can do some uh some fixing.
So that's you know what we call the dashboard experience. What is amazing there is you know if I use Salesforce every day I don't want to go to hotspot I go to Salesforce and here inside Salesforce I have access to all the liveboard capabilities. I have access to spoiler again. Same experience here. I can ask all my questions. It's really powerful. Again, also if I have like some PowerPoint presentation or Google slide, I can embed spotter, I can embed spot like right away where I work.
So right in PowerPoint I can scroll through all the uh so dashboard and I can plug directly the visualization from there and the really cool features you know like with creating like you know a presentation that you have to update every week on Monday to send to your to your boss right here in one click is going to refresh all the slide with the latest data. So uh really cool and of course you have the same in Google sheet.
So here believe it or not but this is coming right away from satspot. Cool. Now okay so next you are like you know Frana we are in 2025 live is great but what if I want to speak directly to my analysts right? What if I don't want to go to the dashboard? I say sure great let's do that. So here you can use also spotter as a free form chat windows. So here again on the same uh customer churn database model and you know the first question I'm going to ask is actually I don't really know what question can I ask because one of the biggest question in analytics is actually what is the right questions.
So here I just ask what is this data? So what happen when you do that? We are going to query the semantic layers that you have created and inside the semantic layers you have all the columns you have the descriptions they can come from uh unity of course from datab bricks but it's really going to give you like context of this information and then spotter can query directly. So in this example, sporter was able to tell me like actually Frana you can run trend analysis prediction because these are the columns available in your data set and these are some example of questions that you can ask but basic you can also uh do analysis on customer segmentation and demographic or you can do analysis on financial usage and analytics.
So I kind of want to continue in my trend analysis. So I'm going to ask him like can you know what can uh we do about the number one customer churn analysis and here I'm going to go a level a level uh deeper. So here again for customer analysis these are all the columns that you should use with some descriptions and basically by the way this description is not part of our semantic layers. This has been generated automatically.
Then here are some type of question I can ask. So I'm like yeah okay good suggestions let's just go to other questions. So here very simply it's going to ask all the questions using the same search token technologies that I show you before right so you really understand what's what has been done and you can uh you can kind of control it so we support table we support graphic so here the top 10 reason of churn where churn equal true and sorted by churn reason.
So again very easy to understand and if you don't like the answer you can coach the system. So I can say you know what when I ask this question actually I don't want to be treated like that. So you can edit the answer and add it to the coaching so that the system becomes smarter time over time but it's not over right um spotter can do more one thing I really like is actually when is creating like calculation on my behalf.
So here for example I'm like what is a customer lifetime value by state? we in the data set in the model we don't have anywhere the definition of customer item value but customer item value is a well-known uh formula right so spoiler is going to ask to LLM what is the definition of customer item value and then we are going to map it to the tables we have to the semantic layers that we have created and using the semantic layers and definition from LLM we are creating the formula directly on the fly so then I have my result again I I don't need to go somewhere to create a formula.
It's all done uh automatically which is u really cool. The last two thing I want to show you before to move on is uh from spotter is what we call oh it's not this one sorry more skills. So you have to think about spotter like you know it's like a teenager right you are training him to do stuff. So today you can answer data questions, you can answer uh what we call data literacy which is data about your semantic layers.
But we have also added some new other skills. The skills that I share with you on the liveboard understanding why number are changing. We also transporter to answer these type of questions. So now if I go here and I ask him like what is my churn? I'm going to be able to ask a follow-up question which is here my number increase. So why is my number increasing and very easily again using the skills that we call spot IQ which is a change analysis is going to be able to tell me exactly what change.
So again right through a conversation uh experience. So here Franoa churn have increased by 42% not great obviously and these are the different insight you should be aware of and again all the data to back the insight here are available through the same interface so through the same visualization again I can edit I can save I can create a lifeboard from it it's really like an embedded experience sporter is not like on the side but it's really like fully part of our uh of our product and so I show you like you know spotter as your personal assistant I show you like how you can have lightboard with AI embedded how you can embed it in Salesforce how you have it on mobile how you have it on Google drive powerpoint we also can and this is very common use case for our customers how can they can bring all hotspot in their own applications so here it's an example of company Nexora and Nexora have decided to push this information to their customer directly.
So now you know everything like that I show you before it's available in this completely like uh custom application created by fictious customers they can embed all this information as an OEM or you know white label if you say nobody will know it's coming from hotspot but it will all of them without having to build everything to do that really cool so my last demo it's one of the latest things that we have announced actually last week uh it's it's about spotter but you know who is aware of the MCP uh protocol who ever heard about it so not many people so MCP is basically a way for um to bring more information to an agent so as you know like you know in your company you may already have created an agent to answer knowledge questions and you want to be able to bring a new skills to the agent you want this agent that you have created be able to answer data question for example and to be able to do that the MCP is very helpful because very easily you are going to be able to plug to your own agent in your company the skills of spotter and this is what we have released recently so to be able to set up that it's very easy right it's not like API code in every way complex things to do it's the only things that I had to do is specify the link to the MCP server server of hotspot doing that I'm using now uh uh an agent in this case I'm taking cloud because cloud support the MCP protocol and so now cloud can speak directly to hotspot thanks to the line I show you here so you can do that with whatever you know agentic you may have developed in your company and so hotspots will have multiple skills it will be able to create get relevant questions which is kind of accessing our reasoning engine.
It can get straight answer to questions and it can even create dashboard. So now I'm just going to show you an example of that. So here I'm going to ask cloud uh so I want to raise the price of my vehicle. There you go. Next year how should I approach this? So cloud is going to uh look at all the different integration is available to him and so looking at sspot he's going to leverage it and first call the get relevant questions.
So if I want to increase the price what are the sub question I need to ask for sspot is going to answer this questions pretty soon. So it's live as you can see here up. So it provided like all the its of all the question you should ask for and then cloud is going to ask every question one by one to be able to come up to a final uh conclusion of this things. So I don't want to you know we are little bit on time. So let me go back to a previous previous uh questions that I did.
Here we go. So you know through the same process here are the key finding. So all the data to back what code were saying is based on queries that was done dynamically through hotspots and the query was also you know dynamically done on your data bricks uh directly no implementation nothing and then so you bring very valuable insight using this data coming from spot and now you know how to increase uh your car to make more profit next summer so it's kind of cool so that's that's was a presentation about you know sspot a little bit about what we are doing, what are the main capabilities.
Um, so if you want to learn more, uh, we have launched what we call uh, data spot which is a specific edition for data bricks. We have a booth downstairs uh, you can visit us and uh, you can scan this QR code which I don't know what is going to do but you can try and you will see. Thank you very much and again thank you for uh, for your presentations.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.