Informatica World 2024 | Gaurav Pathak | Informatica

Clips
News
More from Informatica World 2024

CLAIRE GPT: How Informatica is transforming enterprise data with AI

As organizations continue to delve deeper into how artificial intelligence can be a game changer, CLAIRE GPT is emerging as a resourceful AI-powered data management tool meant for the enterprise.By using a natural language interface to propel data management and discovery, CLAIRE GPT is changing the odds because it’s ChatGPT for enterprise data, which is grounded in enterprise reality using a metadata system of record, according to Gaurav Pathak (pictured), vice president of product management AI and metadata at Informatica Inc.Informatica’s Gaurav Pathak talks to theCUBE about the importance of CLAIRE GPT

play_circle_outline Empowering Data Management in Informatica World through Generative AI and the Human-in-the-Loop Approach

play_circle_outline Overview of CLAIRE GPT and its role in data wrangling and gen AI

play_circle_outline Unleashing the Power of GPT Integration and Metadata: Empowering Informatica to Revolutionize Enterprises

play_circle_outline Security and privacy concerns and the architecture of CLAIRE GPT

Info

Gaurav Pathak | Informatica

search

Rebecca Knight

>> Hello everyone and welcome back to theCUBE's live coverage of Informatica World here at the Mandalay Bay Convention Center. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Rob Strechay. Rob, we're having great conversations at the show about the transformative impact of generative AI on data management, and in particular the introduction of CLAIRE GPT.

>> Yeah, yeah. I think, again, even though it came out before here and had been in preview, I think what you're seeing is that how it helps to ease the burden for data wrangling and how it really eases the burden for getting to gen AI. And I think that, to me, is a huge advantage for data engineering.

Rebecca Knight

>> Well, here to talk more about that is our next guest. He is Gaurav Pathak, vice president product management, metadata and CLAIRE at Informatica. Thank you so much for returning to theCUBE, Gaurav. I know you and Rob had a good conversation just a little while ago.

>> No, thank you, Rob, and thank you, Rebecca, for having me back.

Rebecca Knight

>> So, as Rob was saying, CLAIRE GPT came out a few weeks ago, but this was really the coming out party here at Informatica World. Why don't you tell our viewers who maybe have missed the keynote a little bit about CLAIRE GPT and what it does?

>> Sure. And CLAIRE GPT for us, like I said on the main stage, has been a labor of love. We have worked very hard on it for the last year. Our vision for AI, for data management has been AI that understand and organization's data assets and data ecosystems completely and is grounded in that reality, and that's what we wanted to do. We have all seen the ChatGPT for internet data, being able to ask all the questions and understanding, creating code, creating poems, creating all sorts of homeworks for us, and we wanted to build something like that, but for enterprise data. ChatGPT for my enterprise data, and that's what CLAIRE GPT is all about. Being able to ask questions like, "I'm doing customer churn analysis. Can you help me find the right data assets?" And when we ask a question like that, we want CLAIRE GPT to provide answers that's based on that particular organization, on what this particular user wants. So, that's what CLAIRE GPT is all about. It's ChatGPT for enterprise data, grounded in enterprise reality with the metadata system of record we built in.

>> Yeah, I think what was amazing to me, and by the way, saying it again, that doing keynotes with demos is always tough and you can have pop-ups and other stuff from IT that we all know about, but you start to look at it and go, okay... One of the things I think was my most important takeaway was the fact that the human in the loop aspect of it because it is your data and you are grounding it, you want to put governance and you want to put guardrails around that. And I think people are always so worried about AI and it's going to do everything, but I think there is an important context. And that was part of the demo as well, was how do you use this as an assistant, not as in a replacement? And I think that's the big thing that I took away from it as well.

>> Definitely. And on the main stage demos, completely agree. Earlier, we used to pray to the demo gods, now we pray to the AI gods as well that they're in the right mood and are able to answer questions properly. Getting AI to understand questions like, "What is the revenue of my Product X in quarter Y?" Being able to understand each of those terms for this organization grounded in, oh, product X means this is what you're talking about, quarters are different for each organization. The revenue, is it subscription revenue, ACV? Every organization does it differently. So, being able to ground it in that kind of business context, finding the right data sets, that's what we wanted to do with CLAIRE GPT.

>> Yeah, I think to just build off that, because in a previous discussion when we're talking about the next intelligent data platform and really metadata's at the core of that. And I think to your point, and we discussed this a little bit, is that how somebody defines their fiscal year, for instance. We were last week with a customer, they're in fiscal year '25 already. It started April 1st. When you start to look at this, if you're not having that grounding and you're not having that metadata, and I think that's really what I've seen is one of the integrations between GPT and the metadata seems to be one of the superpowers of this.

>> And that's one of the unique capabilities of Informatica. Being able to create this metadata system of record for each organization that we work with. As they start with any of our products, we are creating a map of what are the most important data assets and how they are related to each other. That metadata repository, which actually is in a graph database, serves as the repository for CLAIRE GPT for it to understand where the data assets are and answer these kind of questions as well. The problems after people saw those demos, the customers... I've met a lot of customers after that, great interest. But the questions now they're coming up with, "Can it create my data quality rules now that understands my metadata? I just point it to my data set and have it create all data quality rules for it?" For that, we need to have it understand that domain. We need to have it trained on medical textbooks because our customers want to create data quality rules on heart rates and the blood pressures and all of that. It needs to understand that domain data very well. That's what we'll be working on next.

Rebecca Knight

>> Well, and it gets back to something that another guest that we had on the show yesterday said about talking human to your data. You don't need a degree in computer science. You don't need to be a data scientist to be able to ask these questions of CLAIRE GPT. Can you talk us through some of the... I mean you've really pointed out so many of the benefits that this could bring, but there have to be challenges and risks. What are the things that you are thinking about and that really that keep you up at night in terms of making sure that the promise and the potential of CLAIRE GPT is realized?

>> That's a big question. With technology like this, our users expect the world out of CLAIRE GPT. They want us to create AI models of their data. They want us to create data pipelines, data quality rules. They want everything that IDMC does, which is our Informatica Data Management Cloud, just using natural language prompts. And what we want to do is allow CLAIRE GPT to do that responsibly, understanding each organization's datasets very, very well. So, what we have trained CLAIRE GPT to do is that if it is really confident on the answer, based on what it knows from the metadata system of record, only then answer those questions. We have made it really refusal aware, where we have told it that if you don't know the thing that the user is asking about, don't be shy to say no. And that's what users will find with CLAIRE GPT. Given that this is enterprise data, we don't want the model to go awry and start hallucinating and creating biases. Those are not things that we don't want the model to get involved in. So, that's something that keeps me up at night to make sure that the model is aligned with our users' preferences. That's what we are trying to make sure that it continues to stay always aligned with.

>> Yeah, and I think just bridging off of that, I think one of the other things that keeps customers up at night is really the privacy aspect of it. Why don't you talk about that because I think that's important for them to understand as well?

>> Absolutely, and that goes into the system architecture of CLAIRE GPT. I remember when we announced CLAIRE GPT and there were a lot of customers who were interested in it. I remember talking to this healthcare customer, I was going to meet them. And even before I said hi to them or introduced myself to them, the first thing they said is, "If you're sending my data to third-party AI providers, don't even talk to me about it." And that's what has been the case with a lot of our enterprise customers, the Fortune 5, Fortune 50, Fortune 100 customers. They don't want their metadata or data to go to third-party providers, and that's the security and privacy concern. So, our architecture uses open source small language models or efficient language models, if I may call them. There are about eight of them running behind the scenes of CLAIRE GPT. We have fine-tuned them on data management tasks. We have tuned them, oh, this particular model can generate data pipelines. This particular model can convert your natural language to SQL. This particular model can give you lineage queries and so on. Each of these models makes sure that your metadata or customer's metadata does not ever go out of Informatica Cloud. We make sure that we never store any data points within our response database, and we never learn from prompts of the end users as well. Unless somebody says thumbs up, thumbs down. We use that for aligning our models. But other than that, there's no learning, there's no sending of data to any third-party providers. We've always designed for our large customers, and that's our promise to our customers, that we will use responsible AI as we work with CLAIRE GPT.

Rebecca Knight

>> So, Informatica is such a customer-centric company, customer-driven, and you're describing how even though it's just such early days with CLAIRE GPT, customers are already expressing interest. Before you can even get your name out, they're asking you questions about it. Talk a little bit about some of the future trends that you're seeing in terms of what their concerns are and how that's going to impact and affect how you build more functionality into CLAIRE GPT.

>> Sure, and I hear some of those concerns. And then for me, those look like opportunities in designing enterprise software really for making the experience of humans designing data management pipelines better, all the data management jobs better. The security privacy, responsible AI concerns, those are important to make sure that helps in designing CLAIRE GPT to the specifications that our enterprise customers want. What they want more from here is more intelligence, being able to just point to their data state and say, "Oh, make sure that the data quality of this schema is all right," which is a very hard problem because you need to now understand the domain. You need to make sure that everything is correctly set up. Those are things that we are working on. We are working on multi-step reasoning, which is you would've seen all these softwares like Devon and the open source software engineers and so on. We want to create a data management engineer that can work with human in the loop for creating all those complex data management projects as well.

>> Yeah, and I'll give you another bit of credit to the Informatica team that's been on stage and doing the keynotes, especially yesterday when you guys were going through because you have roadmap on there. And by the way, it's the longest slide I've ever seen in my life. But you've got just a mountain of stuff you want to get to. And obviously, roadmaps can change the second you get off stage, all that, good disclosures. But what's exciting you about what's on that roadmap and where you're going to go and the challenges that will get addressed for those customers you're talking to?

>> So, number one on the CLAIRE roadmap is bringing a data management copilot in every single service of IDMC. Whether you are a data engineer in integration, creating data pipelines, a data quality steward creating data quality rules or a governance steward creating business glossary terms or policies. We want gen AI to be helping you everywhere in every single layer, every single service of IDMC. That's something that we are really excited about. We are showing it to a lot of customers at Informatica World, and we'll be working with a lot of them to get these out towards... A lot of it's coming this year. We are also making data integration and application integration capabilities working through natural language copilots. One of the things you would've heard, Rob, in the main stage keynote yesterday was all the IDMC for gen AI capabilities, which is being able to take unstructured content, chunk it, embed it into vector databases, and then enable gen AI use cases for customers as well. And those are other areas that we are investing in as well.

Rebecca Knight

>> Yes, I was also struck by the presenting of the product roadmap, which is not something you see at a lot of these shows. Finally, I would love you to just reflect on this moment in time that we're living in. We're really at this inflection point in terms of the new era of generative AI, and it's so exciting and has so much promise and potential, but there are still so many risks, particularly when it comes to the foundation of the data. I mean, how do you reconcile the excitement on one hand with the real challenge before you on the other, and in terms of your conversations with customers too, how do you temper your optimism with also really, okay, this is great, but we got to make sure all these boxes are checked first?

>> Absolutely. And then, I think that's really the party trick here, being able to balance both of those, being able to balance the excitement around this gen AI innovation. I remember me first trying out ChatGPT when it came out in November of 2022, and it definitely felt that this is a moment of change. This is not something where we have had software do anything like this. Being able to understand really natural language in deep, even as you go in deeper into context, into other different domains, it was able to understand you and was able to follow you. So, being able to use all of that innovation and then make sure that all the enterprise guardrails and security, privacy, grounding in enterprise data asset reality is followed as well. Trying to balance both of these is going to be really the trick. The way we are trying to do it is as we work with customers, we are trying to train our metadata models based on what they are seeing today, what their data management users want today, and then bringing in all these new capabilities like multi-step reasoning from gen AI, being able to even look at multi-modal stuff that we saw, video and audio, and then being able to use those modes to make data management easier. In the short term, I think what that means is it makes our data management users a lot more productive. It makes accessibility to data a lot more easier. We've forever talked about data democratization, and I think finally we are getting close to realizing it because business users can ask these questions in their language and start getting answers of data. In the long term. I think we are going to see very interesting use cases. If it is so easy generating data management code, what does that mean to maintain it and reuse it, and then what happens should we all generate it at the time when it is needed? Should we have AI decide data management rules for us? That will be the thing.

>> Yeah, I mean that to me, and I can't wait to have the next conversation with you because to me it also talks about, hey, you're building these data products using certain metadata and it was generated by AI. And now, I need to go back and actually edit it. And how am I going to edit it? Am I going to edit it with the AI again or... There's so many things that could happen.

Rebecca Knight

>> So many good questions. Excellent.

>> That's right. And then, today we don't think you can generate a full pipeline using natural language, so that's why we create the pipeline, but allow people to tweak it and make it really perfect for their enterprise use cases. I think we'll stay in this state for a while and then we'll wait for the next leap in gen AI to make full pipelines possible.

Rebecca Knight

>> Excellent. Well, thank you so much for coming on. Always great insights from you, Gaurav.

>> Thank you, Rebecca. Thanks, Rob, for having me again.

>> Yes.

Rebecca Knight

>> I'm Rebecca Knight for Rob Strechay. Stay tuned for more of theCUBE's live coverage of Informatica World. You're watching theCUBE, the leader in enterprise tech news and analysis.

Gaurav Pathak | Informatica

search

Rebecca Knight

>> Hello everyone and welcome back to theCUBE's live coverage of Informatica World here at the Mandalay Bay Convention Center. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Rob Strechay. Rob, we're having great conversations at the show about the transformative impact of generative AI on data management, and in particular the introduction of CLAIRE GPT.

>> Yeah, yeah. I think, again, even though it came out before here and had been in preview, I think what you're seeing is that how it helps to ease the burden for data wrangling and how it really eases the burden for getting to gen AI. And I think that, to me, is a huge advantage for data engineering.

Rebecca Knight

>> Well, here to talk more about that is our next guest. He is Gaurav Pathak, vice president product management, metadata and CLAIRE at Informatica. Thank you so much for returning to theCUBE, Gaurav. I know you and Rob had a good conversation just a little while ago.

>> No, thank you, Rob, and thank you, Rebecca, for having me back.

Rebecca Knight

>> So, as Rob was saying, CLAIRE GPT came out a few weeks ago, but this was really the coming out party here at Informatica World. Why don't you tell our viewers who maybe have missed the keynote a little bit about CLAIRE GPT and what it does?

>> Sure. And CLAIRE GPT for us, like I said on the main stage, has been a labor of love. We have worked very hard on it for the last year. Our vision for AI, for data management has been AI that understand and organization's data assets and data ecosystems completely and is grounded in that reality, and that's what we wanted to do. We have all seen the ChatGPT for internet data, being able to ask all the questions and understanding, creating code, creating poems, creating all sorts of homeworks for us, and we wanted to build something like that, but for enterprise data. ChatGPT for my enterprise data, and that's what CLAIRE GPT is all about. Being able to ask questions like, "I'm doing customer churn analysis. Can you help me find the right data assets?" And when we ask a question like that, we want CLAIRE GPT to provide answers that's based on that particular organization, on what this particular user wants. So, that's what CLAIRE GPT is all about. It's ChatGPT for enterprise data, grounded in enterprise reality with the metadata system of record we built in.

>> Yeah, I think what was amazing to me, and by the way, saying it again, that doing keynotes with demos is always tough and you can have pop-ups and other stuff from IT that we all know about, but you start to look at it and go, okay... One of the things I think was my most important takeaway was the fact that the human in the loop aspect of it because it is your data and you are grounding it, you want to put governance and you want to put guardrails around that. And I think people are always so worried about AI and it's going to do everything, but I think there is an important context. And that was part of the demo as well, was how do you use this as an assistant, not as in a replacement? And I think that's the big thing that I took away from it as well.

>> Definitely. And on the main stage demos, completely agree. Earlier, we used to pray to the demo gods, now we pray to the AI gods as well that they're in the right mood and are able to answer questions properly. Getting AI to understand questions like, "What is the revenue of my Product X in quarter Y?" Being able to understand each of those terms for this organization grounded in, oh, product X means this is what you're talking about, quarters are different for each organization. The revenue, is it subscription revenue, ACV? Every organization does it differently. So, being able to ground it in that kind of business context, finding the right data sets, that's what we wanted to do with CLAIRE GPT.

>> Yeah, I think to just build off that, because in a previous discussion when we're talking about the next intelligent data platform and really metadata's at the core of that. And I think to your point, and we discussed this a little bit, is that how somebody defines their fiscal year, for instance. We were last week with a customer, they're in fiscal year '25 already. It started April 1st. When you start to look at this, if you're not having that grounding and you're not having that metadata, and I think that's really what I've seen is one of the integrations between GPT and the metadata seems to be one of the superpowers of this.

>> And that's one of the unique capabilities of Informatica. Being able to create this metadata system of record for each organization that we work with. As they start with any of our products, we are creating a map of what are the most important data assets and how they are related to each other. That metadata repository, which actually is in a graph database, serves as the repository for CLAIRE GPT for it to understand where the data assets are and answer these kind of questions as well. The problems after people saw those demos, the customers... I've met a lot of customers after that, great interest. But the questions now they're coming up with, "Can it create my data quality rules now that understands my metadata? I just point it to my data set and have it create all data quality rules for it?" For that, we need to have it understand that domain. We need to have it trained on medical textbooks because our customers want to create data quality rules on heart rates and the blood pressures and all of that. It needs to understand that domain data very well. That's what we'll be working on next.

Rebecca Knight

>> Well, and it gets back to something that another guest that we had on the show yesterday said about talking human to your data. You don't need a degree in computer science. You don't need to be a data scientist to be able to ask these questions of CLAIRE GPT. Can you talk us through some of the... I mean you've really pointed out so many of the benefits that this could bring, but there have to be challenges and risks. What are the things that you are thinking about and that really that keep you up at night in terms of making sure that the promise and the potential of CLAIRE GPT is realized?

>> That's a big question. With technology like this, our users expect the world out of CLAIRE GPT. They want us to create AI models of their data. They want us to create data pipelines, data quality rules. They want everything that IDMC does, which is our Informatica Data Management Cloud, just using natural language prompts. And what we want to do is allow CLAIRE GPT to do that responsibly, understanding each organization's datasets very, very well. So, what we have trained CLAIRE GPT to do is that if it is really confident on the answer, based on what it knows from the metadata system of record, only then answer those questions. We have made it really refusal aware, where we have told it that if you don't know the thing that the user is asking about, don't be shy to say no. And that's what users will find with CLAIRE GPT. Given that this is enterprise data, we don't want the model to go awry and start hallucinating and creating biases. Those are not things that we don't want the model to get involved in. So, that's something that keeps me up at night to make sure that the model is aligned with our users' preferences. That's what we are trying to make sure that it continues to stay always aligned with.

>> Yeah, and I think just bridging off of that, I think one of the other things that keeps customers up at night is really the privacy aspect of it. Why don't you talk about that because I think that's important for them to understand as well?

>> Absolutely, and that goes into the system architecture of CLAIRE GPT. I remember when we announced CLAIRE GPT and there were a lot of customers who were interested in it. I remember talking to this healthcare customer, I was going to meet them. And even before I said hi to them or introduced myself to them, the first thing they said is, "If you're sending my data to third-party AI providers, don't even talk to me about it." And that's what has been the case with a lot of our enterprise customers, the Fortune 5, Fortune 50, Fortune 100 customers. They don't want their metadata or data to go to third-party providers, and that's the security and privacy concern. So, our architecture uses open source small language models or efficient language models, if I may call them. There are about eight of them running behind the scenes of CLAIRE GPT. We have fine-tuned them on data management tasks. We have tuned them, oh, this particular model can generate data pipelines. This particular model can convert your natural language to SQL. This particular model can give you lineage queries and so on. Each of these models makes sure that your metadata or customer's metadata does not ever go out of Informatica Cloud. We make sure that we never store any data points within our response database, and we never learn from prompts of the end users as well. Unless somebody says thumbs up, thumbs down. We use that for aligning our models. But other than that, there's no learning, there's no sending of data to any third-party providers. We've always designed for our large customers, and that's our promise to our customers, that we will use responsible AI as we work with CLAIRE GPT.

Rebecca Knight

>> So, Informatica is such a customer-centric company, customer-driven, and you're describing how even though it's just such early days with CLAIRE GPT, customers are already expressing interest. Before you can even get your name out, they're asking you questions about it. Talk a little bit about some of the future trends that you're seeing in terms of what their concerns are and how that's going to impact and affect how you build more functionality into CLAIRE GPT.

>> Sure, and I hear some of those concerns. And then for me, those look like opportunities in designing enterprise software really for making the experience of humans designing data management pipelines better, all the data management jobs better. The security privacy, responsible AI concerns, those are important to make sure that helps in designing CLAIRE GPT to the specifications that our enterprise customers want. What they want more from here is more intelligence, being able to just point to their data state and say, "Oh, make sure that the data quality of this schema is all right," which is a very hard problem because you need to now understand the domain. You need to make sure that everything is correctly set up. Those are things that we are working on. We are working on multi-step reasoning, which is you would've seen all these softwares like Devon and the open source software engineers and so on. We want to create a data management engineer that can work with human in the loop for creating all those complex data management projects as well.

>> Yeah, and I'll give you another bit of credit to the Informatica team that's been on stage and doing the keynotes, especially yesterday when you guys were going through because you have roadmap on there. And by the way, it's the longest slide I've ever seen in my life. But you've got just a mountain of stuff you want to get to. And obviously, roadmaps can change the second you get off stage, all that, good disclosures. But what's exciting you about what's on that roadmap and where you're going to go and the challenges that will get addressed for those customers you're talking to?

>> So, number one on the CLAIRE roadmap is bringing a data management copilot in every single service of IDMC. Whether you are a data engineer in integration, creating data pipelines, a data quality steward creating data quality rules or a governance steward creating business glossary terms or policies. We want gen AI to be helping you everywhere in every single layer, every single service of IDMC. That's something that we are really excited about. We are showing it to a lot of customers at Informatica World, and we'll be working with a lot of them to get these out towards... A lot of it's coming this year. We are also making data integration and application integration capabilities working through natural language copilots. One of the things you would've heard, Rob, in the main stage keynote yesterday was all the IDMC for gen AI capabilities, which is being able to take unstructured content, chunk it, embed it into vector databases, and then enable gen AI use cases for customers as well. And those are other areas that we are investing in as well.

Rebecca Knight

>> Yes, I was also struck by the presenting of the product roadmap, which is not something you see at a lot of these shows. Finally, I would love you to just reflect on this moment in time that we're living in. We're really at this inflection point in terms of the new era of generative AI, and it's so exciting and has so much promise and potential, but there are still so many risks, particularly when it comes to the foundation of the data. I mean, how do you reconcile the excitement on one hand with the real challenge before you on the other, and in terms of your conversations with customers too, how do you temper your optimism with also really, okay, this is great, but we got to make sure all these boxes are checked first?

>> Absolutely. And then, I think that's really the party trick here, being able to balance both of those, being able to balance the excitement around this gen AI innovation. I remember me first trying out ChatGPT when it came out in November of 2022, and it definitely felt that this is a moment of change. This is not something where we have had software do anything like this. Being able to understand really natural language in deep, even as you go in deeper into context, into other different domains, it was able to understand you and was able to follow you. So, being able to use all of that innovation and then make sure that all the enterprise guardrails and security, privacy, grounding in enterprise data asset reality is followed as well. Trying to balance both of these is going to be really the trick. The way we are trying to do it is as we work with customers, we are trying to train our metadata models based on what they are seeing today, what their data management users want today, and then bringing in all these new capabilities like multi-step reasoning from gen AI, being able to even look at multi-modal stuff that we saw, video and audio, and then being able to use those modes to make data management easier. In the short term, I think what that means is it makes our data management users a lot more productive. It makes accessibility to data a lot more easier. We've forever talked about data democratization, and I think finally we are getting close to realizing it because business users can ask these questions in their language and start getting answers of data. In the long term. I think we are going to see very interesting use cases. If it is so easy generating data management code, what does that mean to maintain it and reuse it, and then what happens should we all generate it at the time when it is needed? Should we have AI decide data management rules for us? That will be the thing.

>> Yeah, I mean that to me, and I can't wait to have the next conversation with you because to me it also talks about, hey, you're building these data products using certain metadata and it was generated by AI. And now, I need to go back and actually edit it. And how am I going to edit it? Am I going to edit it with the AI again or... There's so many things that could happen.

Rebecca Knight

>> So many good questions. Excellent.

>> That's right. And then, today we don't think you can generate a full pipeline using natural language, so that's why we create the pipeline, but allow people to tweak it and make it really perfect for their enterprise use cases. I think we'll stay in this state for a while and then we'll wait for the next leap in gen AI to make full pipelines possible.

Rebecca Knight

>> Excellent. Well, thank you so much for coming on. Always great insights from you, Gaurav.

>> Thank you, Rebecca. Thanks, Rob, for having me again.

>> Yes.

Rebecca Knight

>> I'm Rebecca Knight for Rob Strechay. Stay tuned for more of theCUBE's live coverage of Informatica World. You're watching theCUBE, the leader in enterprise tech news and analysis.