Informatica World 2024 | Mo Mirza | OMERS & Sumeet Agrawal | Informatica

Clips
News
More from Informatica World 2024

Mohammad Mirza

Director, Enterprise Data, Software & Integrations Engineering

OMERS

Sumeet Agrawal

Vice President of Product Management

Informatica

OMERS and Informatica forge path for enterprise innovation through gen AI integration

As artificial intelligence continues to call the shots in the enterprise world, gen AI integration has become a burning issue based on the need to augment human capacity.With Ontario Municipal Employees Retirement System being one of the largest pension firms on Canadian soil, Informatica Inc. emerged as a perfect choice in the realization of enterprise-class gen AI applications because the company’s policy is that every app ought to meet strict trusted AI, governance and security guidelines, according to Mo Mirza (pictured, left), director of engineering at OMERS. OMERS’s Mo Mirza and Informatica’s Sumeet Kumar Agrawal talk to theCUBE about the importance of gen AI integration

pause_circle_outline Unveiling Informatica World: Embracing Data Governance for Optimal Performance, Featuring OMERS' Sizable Pension Data

play_circle_outline The Crucial Collaboration: Enhancing AI Efficiency through Engineering, Data Science, and OMERS' Expertise

play_circle_outline Use of Informatica's platform to decouple chat chat's API, allowing for flexibility and scalability

play_circle_outline OMERS and Informatica: Strengthening Partnership for GenAI, Integration, and Scalability in an Agile Enterprise

Info

Mo Mirza | OMERS & Sumeet Agrawal | Informatica

Mohammad Mirza

Director, Enterprise Data, Software & Integrations Engineering OMERS

Sumeet Agrawal

Vice President of Product Management Informatica

search

>> Hello everyone and welcome back to theCUBE's live coverage of Informatica World here at the Mandalay Bay Convention Center in Las Vegas. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Rob Strechay. GenAI word of the day.

>> Absolutely. And I think making it real and governed and bringing policy to it really is a big piece of that where how do you do it and how do you get moving with it? All the customers are here talking about it and I am so excited to continue that conversation.

>> To get their data ready for it.

>> Yes.

>> Exactly, exactly. Well, with that, I'd like to welcome our next two guests to the show. We have Mo Mirza, he is the Director, Enterprise Data Software and Integrations Engineering at OMERS. Welcome Mo.

>> Thank you.

>> And Sumeet Agrawal, VP of product Management at Informatica. Thank you so much.

>> Very excited to be here.

>> Yes, and great job on the keynote note this morning.

>> Thank you. Thank you.

>> It was a brilliant presentation. Mo, I'm going to start with you. Why don't you tell our viewers a little bit about OMERS, what you're all about?

>> Sure, absolutely. At OMERS, we're a defined benefit pension plan covering some most important people in the province of Ontario. We have 620,000 plan members covering city workers and police and fire and emergency medical services workers. We're an organization that is contributing to our Ontario GDP and so 1 in every 11 households in Ontario depends on an income from OMERS.

>> So I'm imagining you have a lot of data that you're dealing with and a lot of data challenges.

>> Yeah, we're also one of Canada's largest institutional investors, and so we have a lot of data that is really operational and my purview is really around the operational data and the platforms that we run at OMERS. I think it's important to also understand that for how we run our engineering teams and our platform engineering, we work very closely with many of our data scientists as well as our artificial intelligence experts in our COE. And so we work as a service provider back to them.

>> And I think again, it's especially where you have such a public interest in your corporation and you're doing investments and things, it's pretty critical, like you said, not only to the economy but to those people. And now being able to be customer focused and help them out, I'm sure there's a lot that is going on with that data, but there's also the GenAI component to this and how GenAI is injecting itself into that. How do you see that playing out with what you're dealing with?

>> So from an engineering perspective, there are no shortcuts around this. And so for us, we're not the experts in GenAI, but we look at it as an embedded component of a data pipeline or another integration tool in our toolkit or a feature within an application. People want to box it in one way or another, but we look at it as all of those. And so for us, a strategy around the deployment of that, working with the experts around our AI COE is doing it at scale using an integration platform is really kind of an important part of it. I said there's no shortcuts around this. And so, one of the things that, proud to say, is that our organization with the experts that are in the AI COE put a trusted AI guidelines together and that gave us the green light for on the engineering side to be start to building those solutions so that we can take them to market.

>> So Sumeet, I want to bring you in here. How do you work with OMERS and in terms of the use cases that MO is talking about? How do you partner with...

>> No, absolutely. So first of all, thanks Mo for joining us at Informatica World and actually Mo spoke about his use case yesterday in our market perspective session. And I've been associated with OMERS from few months and they're doing some very exciting things and they have been a long-standing customers for our IDMC and especially they're using our API and app integration product lines. So my role here is to support OMERS for their use cases, there are a lot of interaction between us. And when I heard about Mo's vision of making their chatbot more GenAI and he shared how he's using our product lines to do that, that was simply very exciting. So maybe, Mo, if you want to share how you got our integration users who are using our CAI, how you got them converted into GenAI.

>> Yeah, again, as I said, there's really from an integration engineering perspective, another tool in the toolkit for us. So I think there is a level of intimidation that's there with respect to the hype. I think we're seeing users of these technologies get that immediate gratification with that interactive experience for getting insights out of the data that they're using. The approach we took around that is, again, we have a grounding on the integration platform and the integration platform provides us application integration capabilities. The GenAI services that we're seeing there being offered are really no different than any other SaaS offering that's out there. And so it fits in nicely when we are doing a solution to be able to do that. So when we're doing the development, one of the first things we look at is does it fit our integration pattern? And so then if you've got an integration platform and you've got the quick start items that needed to be able to do that, a trusted AI guideline of being one important part of that, then the engineering team has an ability to be able to do that. I think, Sumeet, the partnership we've had with Informatica has been fantastic. I mean going and working directly with the product team, providing them the feedback that we need to be able to support the use cases that are coming through. We've got a lot of excitement on our business side. We've got some fantastically smart people in our data and technology space who are coming to us with suggestions around how we can deploy this or what needs to be deployed. Our engineering team is working really hard to be able to deploy it and scale and integration is part of that.

>> And I would imagine that, again, it's using different parts of the Informatica toolkit to actually get to a trusted AI platform because again, it goes back to you're dealing with financial data, a lot of PII, there's a lot of regulations in Canada similar to GDPR and things of that nature. How are you using the different parts of Informatica toolkit to actually affect that?

>> So as I said, we're looking at the cloud application integration capability, and so if I can speak directly to kind of our use case around that, one of the challenges we saw is that with the rapid adoption of the public chat applications, we wanted to be able to, working with the business and the experts within our data and technology team, enable the rest of our organization to be able to use those tools and take advantage of it. What we did is we created, the software engineering team developed an application working with the business to enable that same sort of chat experience but on our internal cloud, on our private cloud. That helps protect our data. It helps to incentivize good data practices around leakage as well as what's sharing and going outside of our bounds there. Where the integration platform comes into play to be able to now decouple what is any one sort of chat chat's API that's being used in the backend. And so for us when we deploy an application like that, we feel comfortable that if there is a request for us to scale that out for a line of business, add in greater logic in there, or if we need to change, if something happens in the market and an LLM provider is now perhaps falling out of disfavor or we need to actually augment that with another LLM service provider, it's a matter of using a service connector and changing the LLM in the backend and the user experience still remains the same.

>> So I want to amplify what Mo just said. So one of the thing we actually announced at Informatica World, this conference, is our no-code GenAI app and OMERS is one of our great users for that. But what Mo just described is that we actually provide a universal connectivity, which is super easy for any users to just come in and build any LLM connector. So he's talked about service connector, that's one of the way for you to bring any LLM connector, and that's what they have done where they have just used API from LLMs and they were able to build that connector and they were able to orchestrate that in their existing business process and they were able to get that with their own integration developers and they were able to do GenAI. So this is a classic example of how an integration developers can be a GenAI developers.

>> Yeah, I go back to the keynote, great demonstration. So if you haven't seen it again, I would go back and watch it when it's up on the website because it showed how you could do that. It went through, okay, here's how you connect to the vector database.

>> Exactly.

>> Here's how you bring in the LLM. And to your point, I think maybe I want to use Llama 3 now that it's out or I want to go use ChatGPT and 4.0 or whatever that new one is of it and things like that and in doing testing and stuff, has that been a big piece of it as well as going through and iterating through? I would have to say that this has to help with that part of it.

>> It's a very fast moving space right now and it's like almost on a daily basis you'll see either an open source model changing or a new service provider. I think we heard some staggering statistics around how many thousands of new startups there are around these services. For us, what was important, again from an engineering perspective, is we want to be able to respond as quickly as we can to the business requests that are out there. We're by no means experts in LLM services. I don't necessarily know what the fit for purpose is based on the problem solution space for an LLM, but our data science and AI folks do, or business leaders might have some opinion on that. What they're looking from the engineering team is, can you get this integration up and running for us quickly? And so for us to be able to scale this out, we want to be able to take an approach of we've already invested in platforms, we've already invested in technology, how do we get the most out of it for being able to meet the objectives with these new technologies that are coming out?

>> So up on the main stage this morning, Sumeet, you did lay out before the audience the product roadmap. And so I'm curious to hear how you are working with OMERS to determine their roadmap for their GenAI initiatives.

>> Absolutely. So in fact, I want to lay out the scenario that MO have just mentioned that he has taken an integration users who have known how to build Informatica mappings. Now the same users are now doing GenAI development just by a few clicks. Now the next thing is how we can make this enterprise class. Now there are a lot of data science tools out, there are a lot of LLM companies out there and people do use these LLM companies as a science tool. Now where what Informatica is so unique is how we are making it enterprise class. And there are a lot of investment in our road map also, which are happening in this area, which is including API management, building more APIs recipes and all. And maybe Mo, of you want to share how you made this whole process enterprise class at OMERS because that was one of your key bending where what I heard from you before was that you don't want to take any product which is not enterprise class. So tell me more about how you use Informatica to make it enterprise class, the whole GenAI apps.

>> I think again, going back to the integration platform strategy, I think to paraphrase Mike Tyson, everybody's got an AI plan until you crunch the data, right? So what we saw-

>> To paraphrase.

>> Yeah, I paraphrased. So for us, the strategy around AI is sort of evolving and I think to Sumeet's point, what we started developing with our integration muscle was connecting SaaS applications together to be able to operationalize the business and create interoperability. And I think for us, where Informatica's platform has sort of factored into our strategic advantage is around that scalability for the organization around the GenAI applications that might be. So that might be GenAI within a data pipeline so that we can maybe enrich the data or GenAI within an interactive human in the loop sort of application and we can orchestrate that between one or more LLMs as we go along. I think that obviously the sky's the limit in terms of the number of use cases that we're seeing from an engineering team perspective. We want to be able to provide a service around not only the development of the application, but where those GenAI services can be available. Working closely with our data science and AI teams members who are experts in that area around fit for purpose, we're able to now deploy things that we probably wouldn't be able to deploy otherwise.

>> Yeah, I think that to me is really one of the critical things is that the fit for purpose and looking at that and being able to iterate so quickly. So any parting thoughts or recommendations to people as they kind of go down this path? Like you said, not an expert in it, but you may be further down the path because 80% of organizations, we have some polling data, are still in the actual figure it out and not production stage and they're trying to get to where you guys are, even with a limited use case that they have, some of these other organizations, as they grow and understand the fit for purpose.

>> I think integration and your data platform are a key component to that, but it doesn't mean that you have to solve for all the problems. Even on a limited, if you've identified those key capabilities around application integration, so SaaS to SaaS, your data platform, hosting the data and having it cleansed and ready for use, even a limited use case is going to help you break that boundary that you think is insurmountable right now. Working again closely with... I look at data governance as more of an enablement function, so working through that enablement, that's really I think the key items to be able to get through, break through to the next step. GenAI, again, it can be intimidating, but it's really nothing more than another tool in our toolkit around how we process our data, and I think we need to approach it from that perspective. If you've got an integration engineering team, I mean I'm fortunate to have fantastic people working in an integration engineering team that really is one of the key ingredients for quick deployment and success.

>> Don't be scared. I like that. Great advice.

>> Don't be scared. Don't be scared.

>> Great advice.

>> And have a good partner as well.

>> Yeah, yeah, yeah. Absolutely.

>> Indeed. Indeed. Mo and Sumeet, thank you both so much for coming on theCUBE, a really great conversation.

>> Thank you. Pleasure being here.

>> Thank you so much.

>> Thanks.

>> Thank you.

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

Mo Mirza | OMERS & Sumeet Agrawal | Informatica

search

>> Hello everyone and welcome back to theCUBE's live coverage of Informatica World here at the Mandalay Bay Convention Center in Las Vegas. I'm your host, Rebecca Knight, sitting alongside my co-host and analyst, Rob Strechay. GenAI word of the day.

>> Absolutely. And I think making it real and governed and bringing policy to it really is a big piece of that where how do you do it and how do you get moving with it? All the customers are here talking about it and I am so excited to continue that conversation.

>> To get their data ready for it.

>> Yes.

>> Exactly, exactly. Well, with that, I'd like to welcome our next two guests to the show. We have Mo Mirza, he is the Director, Enterprise Data Software and Integrations Engineering at OMERS. Welcome Mo.

>> Thank you.

>> And Sumeet Agrawal, VP of product Management at Informatica. Thank you so much.

>> Very excited to be here.

>> Yes, and great job on the keynote note this morning.

>> Thank you. Thank you.

>> It was a brilliant presentation. Mo, I'm going to start with you. Why don't you tell our viewers a little bit about OMERS, what you're all about?

>> Sure, absolutely. At OMERS, we're a defined benefit pension plan covering some most important people in the province of Ontario. We have 620,000 plan members covering city workers and police and fire and emergency medical services workers. We're an organization that is contributing to our Ontario GDP and so 1 in every 11 households in Ontario depends on an income from OMERS.

>> So I'm imagining you have a lot of data that you're dealing with and a lot of data challenges.

>> Yeah, we're also one of Canada's largest institutional investors, and so we have a lot of data that is really operational and my purview is really around the operational data and the platforms that we run at OMERS. I think it's important to also understand that for how we run our engineering teams and our platform engineering, we work very closely with many of our data scientists as well as our artificial intelligence experts in our COE. And so we work as a service provider back to them.

>> And I think again, it's especially where you have such a public interest in your corporation and you're doing investments and things, it's pretty critical, like you said, not only to the economy but to those people. And now being able to be customer focused and help them out, I'm sure there's a lot that is going on with that data, but there's also the GenAI component to this and how GenAI is injecting itself into that. How do you see that playing out with what you're dealing with?

>> So from an engineering perspective, there are no shortcuts around this. And so for us, we're not the experts in GenAI, but we look at it as an embedded component of a data pipeline or another integration tool in our toolkit or a feature within an application. People want to box it in one way or another, but we look at it as all of those. And so for us, a strategy around the deployment of that, working with the experts around our AI COE is doing it at scale using an integration platform is really kind of an important part of it. I said there's no shortcuts around this. And so, one of the things that, proud to say, is that our organization with the experts that are in the AI COE put a trusted AI guidelines together and that gave us the green light for on the engineering side to be start to building those solutions so that we can take them to market.

>> So Sumeet, I want to bring you in here. How do you work with OMERS and in terms of the use cases that MO is talking about? How do you partner with...

>> No, absolutely. So first of all, thanks Mo for joining us at Informatica World and actually Mo spoke about his use case yesterday in our market perspective session. And I've been associated with OMERS from few months and they're doing some very exciting things and they have been a long-standing customers for our IDMC and especially they're using our API and app integration product lines. So my role here is to support OMERS for their use cases, there are a lot of interaction between us. And when I heard about Mo's vision of making their chatbot more GenAI and he shared how he's using our product lines to do that, that was simply very exciting. So maybe, Mo, if you want to share how you got our integration users who are using our CAI, how you got them converted into GenAI.

>> Yeah, again, as I said, there's really from an integration engineering perspective, another tool in the toolkit for us. So I think there is a level of intimidation that's there with respect to the hype. I think we're seeing users of these technologies get that immediate gratification with that interactive experience for getting insights out of the data that they're using. The approach we took around that is, again, we have a grounding on the integration platform and the integration platform provides us application integration capabilities. The GenAI services that we're seeing there being offered are really no different than any other SaaS offering that's out there. And so it fits in nicely when we are doing a solution to be able to do that. So when we're doing the development, one of the first things we look at is does it fit our integration pattern? And so then if you've got an integration platform and you've got the quick start items that needed to be able to do that, a trusted AI guideline of being one important part of that, then the engineering team has an ability to be able to do that. I think, Sumeet, the partnership we've had with Informatica has been fantastic. I mean going and working directly with the product team, providing them the feedback that we need to be able to support the use cases that are coming through. We've got a lot of excitement on our business side. We've got some fantastically smart people in our data and technology space who are coming to us with suggestions around how we can deploy this or what needs to be deployed. Our engineering team is working really hard to be able to deploy it and scale and integration is part of that.

>> And I would imagine that, again, it's using different parts of the Informatica toolkit to actually get to a trusted AI platform because again, it goes back to you're dealing with financial data, a lot of PII, there's a lot of regulations in Canada similar to GDPR and things of that nature. How are you using the different parts of Informatica toolkit to actually affect that?

>> So as I said, we're looking at the cloud application integration capability, and so if I can speak directly to kind of our use case around that, one of the challenges we saw is that with the rapid adoption of the public chat applications, we wanted to be able to, working with the business and the experts within our data and technology team, enable the rest of our organization to be able to use those tools and take advantage of it. What we did is we created, the software engineering team developed an application working with the business to enable that same sort of chat experience but on our internal cloud, on our private cloud. That helps protect our data. It helps to incentivize good data practices around leakage as well as what's sharing and going outside of our bounds there. Where the integration platform comes into play to be able to now decouple what is any one sort of chat chat's API that's being used in the backend. And so for us when we deploy an application like that, we feel comfortable that if there is a request for us to scale that out for a line of business, add in greater logic in there, or if we need to change, if something happens in the market and an LLM provider is now perhaps falling out of disfavor or we need to actually augment that with another LLM service provider, it's a matter of using a service connector and changing the LLM in the backend and the user experience still remains the same.

>> So I want to amplify what Mo just said. So one of the thing we actually announced at Informatica World, this conference, is our no-code GenAI app and OMERS is one of our great users for that. But what Mo just described is that we actually provide a universal connectivity, which is super easy for any users to just come in and build any LLM connector. So he's talked about service connector, that's one of the way for you to bring any LLM connector, and that's what they have done where they have just used API from LLMs and they were able to build that connector and they were able to orchestrate that in their existing business process and they were able to get that with their own integration developers and they were able to do GenAI. So this is a classic example of how an integration developers can be a GenAI developers.

>> Yeah, I go back to the keynote, great demonstration. So if you haven't seen it again, I would go back and watch it when it's up on the website because it showed how you could do that. It went through, okay, here's how you connect to the vector database.

>> Exactly.

>> Here's how you bring in the LLM. And to your point, I think maybe I want to use Llama 3 now that it's out or I want to go use ChatGPT and 4.0 or whatever that new one is of it and things like that and in doing testing and stuff, has that been a big piece of it as well as going through and iterating through? I would have to say that this has to help with that part of it.

>> It's a very fast moving space right now and it's like almost on a daily basis you'll see either an open source model changing or a new service provider. I think we heard some staggering statistics around how many thousands of new startups there are around these services. For us, what was important, again from an engineering perspective, is we want to be able to respond as quickly as we can to the business requests that are out there. We're by no means experts in LLM services. I don't necessarily know what the fit for purpose is based on the problem solution space for an LLM, but our data science and AI folks do, or business leaders might have some opinion on that. What they're looking from the engineering team is, can you get this integration up and running for us quickly? And so for us to be able to scale this out, we want to be able to take an approach of we've already invested in platforms, we've already invested in technology, how do we get the most out of it for being able to meet the objectives with these new technologies that are coming out?

>> So up on the main stage this morning, Sumeet, you did lay out before the audience the product roadmap. And so I'm curious to hear how you are working with OMERS to determine their roadmap for their GenAI initiatives.

>> Absolutely. So in fact, I want to lay out the scenario that MO have just mentioned that he has taken an integration users who have known how to build Informatica mappings. Now the same users are now doing GenAI development just by a few clicks. Now the next thing is how we can make this enterprise class. Now there are a lot of data science tools out, there are a lot of LLM companies out there and people do use these LLM companies as a science tool. Now where what Informatica is so unique is how we are making it enterprise class. And there are a lot of investment in our road map also, which are happening in this area, which is including API management, building more APIs recipes and all. And maybe Mo, of you want to share how you made this whole process enterprise class at OMERS because that was one of your key bending where what I heard from you before was that you don't want to take any product which is not enterprise class. So tell me more about how you use Informatica to make it enterprise class, the whole GenAI apps.

>> I think again, going back to the integration platform strategy, I think to paraphrase Mike Tyson, everybody's got an AI plan until you crunch the data, right? So what we saw-

>> To paraphrase.

>> Yeah, I paraphrased. So for us, the strategy around AI is sort of evolving and I think to Sumeet's point, what we started developing with our integration muscle was connecting SaaS applications together to be able to operationalize the business and create interoperability. And I think for us, where Informatica's platform has sort of factored into our strategic advantage is around that scalability for the organization around the GenAI applications that might be. So that might be GenAI within a data pipeline so that we can maybe enrich the data or GenAI within an interactive human in the loop sort of application and we can orchestrate that between one or more LLMs as we go along. I think that obviously the sky's the limit in terms of the number of use cases that we're seeing from an engineering team perspective. We want to be able to provide a service around not only the development of the application, but where those GenAI services can be available. Working closely with our data science and AI teams members who are experts in that area around fit for purpose, we're able to now deploy things that we probably wouldn't be able to deploy otherwise.

>> Yeah, I think that to me is really one of the critical things is that the fit for purpose and looking at that and being able to iterate so quickly. So any parting thoughts or recommendations to people as they kind of go down this path? Like you said, not an expert in it, but you may be further down the path because 80% of organizations, we have some polling data, are still in the actual figure it out and not production stage and they're trying to get to where you guys are, even with a limited use case that they have, some of these other organizations, as they grow and understand the fit for purpose.

>> I think integration and your data platform are a key component to that, but it doesn't mean that you have to solve for all the problems. Even on a limited, if you've identified those key capabilities around application integration, so SaaS to SaaS, your data platform, hosting the data and having it cleansed and ready for use, even a limited use case is going to help you break that boundary that you think is insurmountable right now. Working again closely with... I look at data governance as more of an enablement function, so working through that enablement, that's really I think the key items to be able to get through, break through to the next step. GenAI, again, it can be intimidating, but it's really nothing more than another tool in our toolkit around how we process our data, and I think we need to approach it from that perspective. If you've got an integration engineering team, I mean I'm fortunate to have fantastic people working in an integration engineering team that really is one of the key ingredients for quick deployment and success.

>> Don't be scared. I like that. Great advice.

>> Don't be scared. Don't be scared.

>> Great advice.

>> And have a good partner as well.

>> Yeah, yeah, yeah. Absolutely.

>> Indeed. Indeed. Mo and Sumeet, thank you both so much for coming on theCUBE, a really great conversation.

>> Thank you. Pleasure being here.

>> Thank you so much.

>> Thanks.

>> Thank you.

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