Expert Insights on Leveraging Data Governance for Successful AI Implementation

Author: Sue Pittacora

In a recent CDO magazine interview, Sue Pittacora, Chief Strategy Officer at Wavicle Data Solutions, and Sarang Bapat, Director of Data Governance at Mitsubishi Electric Trane US (METUS), sat down for an in-depth discussion about how organizations are navigating the complexities of AI implementation to maximize its value responsibly and effectively.


Throughout the discussion, you can explore how to responsibly implement AI into your organization by integrating a robust data and governance strategy.


Watch the full interview here or scroll down for a detailed transcript of Sarang’s insights. 



Sue: Hello and welcome to the CDO Magazine interview series. I am Sue Pittacora with Wavicle Data Solutions, and I’m delighted to be joined here today by Sarang Bapat, the Data Governance Director at Mitsubishi Electric Trane US, otherwise known as METUS. Welcome Sarang.  


Sarang: Thank you, Sue.  


Sue: So Sarang, we have a great topic today. We’re going to be talking about how organizations are striving to address the challenges and opportunities as they try to harness the value of AI responsibly and effectively.


How did you structure the data governance strategy of METUS? 

Sarang: Thanks for the question. I was drawn to the opportunity to take this role at METUS because this was truly a greenfield opportunity. So, before I joined METUS, we didn’t really have any data governance practice dedicated for data management and data governance within METUS.


If you think about our business, we are a sales and distribution-oriented organization. So, from our perspective, it’s very important to have quality, integrity, and timeliness of our data across order management, systems, distribution, and warehousing. When something goes wrong, things come back to our customer care call centers. So, having that information integrated, timely, and of quality is paramount.  


On top of that, we are also building our analytics capabilities, which means that’s where we are bringing our base business systems data into our analytics warehouse to create insights. So, for that to work correctly, we should be able to get better insights depending upon what quality of data we have. Ownership, lineage, and all of these data governance aspects are important.  


When I started diving into our data governance program, before talking about anything related to the technology, there were two things that I was trying to establish more and more from day one. First is data governance is a business-owned, IT-enabled initiative. It’s not an IT exercise. This is not something that happens in a silo. This is a business-owned, IT-enabled exercise. Number two, data governance is not a project, so it doesn’t have a start and end date. This is going to be an ongoing theme, and it’s rather a strategic lever that we need to effectively use to change our culture to make it a data-first culture. 


So, I talked a lot to our business partners upfront about their challenges, what they are facing, what’s important to the business. Forgetting technology and all other things: what’s important for the business? What are they facing in terms of quality of the data? What happens when it’s not right? What are they doing? How are they reworking and spending a ton of time on productivity and all that?  


So  after that, when we are talking about supply chain domain or sales and distribution, one of the key data sets is usually material master data. Very quickly, that material master becomes like 20,000 parts or pieces of equipment that you sell, and those 20,000 pieces of equipment become 20,000,000 deliveries, and it becomes at scale very quickly and expands into the business pretty much every single day. So, if the information coming out of those 20,000 is not right from a quality perspective, imagine all the downstream issues you’re going to face in your order management, warehousing, customer care, people not receiving things when they’re supposed to, and all those challenges. 


That was the first use case we tried to tackle using our data governance use case scenario. First material master, and then the other one was marketing. The approach was to first talk to the business, understand the problems, and then identify a couple of use cases that we want to dive into. 


I think the approach also is along the lines of not boiling the ocean, but starting something that is meaningful. Then, not drive into all the pillars of governance. Let’s start with basics like metadata management, data quality, and data ownership. Because unlocking that is going to drive a lot more things, like later on, it could be retention, data, issue management, and a ton of different pillars that you can tackle. But these are like the bare bones we need to go after first.  


And then, because it was a greenfield, I needed to work on the people, process, and technology components of the program. Since I was talking to the business early, getting people to buy into the concept of why you should be a data steward for a certain thing was not that difficult. I was able to get them into that thought process and was able to get them to work with me on becoming a data steward for a system.  


Then in parallel, we worked with identifying the vendors, and what’s the better data governance tool, and selected that. We also hired a data governance engineer who would actually implement the data governance platform. 


I’m happy to say after 12+ months of all these exercises, we have six business units on board with around 12 data stewards in place and around 200 critical data elements identified with 250+ data quality rules running currently. So, I’m very excited about our progress within a year. It’s far and wide, but I’m excited for the future.  


Also, the last point I would make is that to have these initiatives going long term, one of the key pieces required is executive sponsorship. We have also set up a data governance council, which is a C-suite institution. Every eight weeks or so, we give them an idea of how the data governance program is going, what progress we are making, even do data stewards spotlights periodically. They talk about their wins and things like that. So, it’s really helping them gain visibility and helping our senior leadership understand what it all means. I’m excited about the progress we have made so far.   


Sue: Excellent. Well, I love how you talk about making data governance business led, and focusing on the business cases, and getting the people from the business side involved versus making it an IT project. And we all know it’s not a project, as you clearly pointed out, it’s an ongoing journey.


How can organizations approach data governance and incorporate people into newer data management models like data mesh and federated data governance? What does your data stewardship model look like? 

Sarang: When you talk about data mesh, or data fabric, or new architectural models, they kind of talk to a point of whether it is a fair model in the sense of are you able to find the data? Are you able to interoperably use the data? Is the data accessible? And can you reuse that model? That’s the fair model that we talk about in data mesh.  


But one of the themes there is that you’re specializing different data sets, like finance data, customer data, or specific to deposits data and cards data in the banking environment. So, you have different federated teams actually creating their own data models, data assets, access requirements, and all of that. And then, if required, put something like a layer that can bridge the gap across systems and things like that – this is one approach.  


But considering the size of METUS and how our business operates, what we have for data governance is a central data governance group. So our responsibility there is mainly creating training, best practices, and what governance means, and then creating the data governance platform and maintaining it. Then we’ll execute that in a hub and spoke model.   


A hub is our data governance piece, which is a center of enablement or excellence. And then on the spokes, you sit in the functional areas for data stewards for order management, supply chain, customer care, warranty pieces, marketing, and things of that nature. 


But their role in our instance is mainly letting us know what’s business critical, providing that metadata and what rules they want to apply. But development of the rules and putting that into the platform is done in a central team, which is my team within IT. But their role is more ongoing monitoring of their assets, data quality, coming up with better rules, being involved in providing access to their assets, and things like that.  


Our model is a little bit more federated – hub and spoke model with a central piece governed by the team that creates the platform and technology. Basically, we add the content and the context, but they tell us what content and context to add. I will put it that way. That’s the model that has been working well for us.


How to develop the right talent strategy to ensure you have the right support for data governance and key stakeholders? 

Sarang: I think this is becoming more and more important. So, data governance before data stewards used to be an add-on responsibility on somebody that already has a day job. So this is like, I may do something or I may not do something with data governance. I’m not hired as a data steward. But believe it or not, that’s changing drastically.  


In my opinion, people who are already in the business who are SMEs – let’s take an example of finance because it’s easy to understand. Some people are very familiar with financial reporting, CCAR reporting in banking, capital calculations, and what they submit to the FED. So, they are very knowledgeable about their subject matter, and at the same time, they use some sort of technology. Maybe a SBase, Oracle, or some sort of reporting, but they are not risk averse towards technology or afraid to use it or daunted by it. I would focus on those kinds of folks who have hybrid skillsets. They have SME knowledge, they are already working with their data, and they use some sort of tech. Teaching a new technology is not that difficult, but having the right mindset about it and not being afraid of it are the kinds of things I would look for in selecting a data steward.  


Then, many times when you’re rolling out data stewardship, you’ll hear people say, “I’m not a data steward.” And then when you explain what a data steward does, they will be like, “I’m actually doing a lot of that.” And like, “exactly, you’re already doing that, and we are just formalizing it and giving you better tools, so you don’t spend a whole lot of time on it.”  


So that makes the sale. It’s not just that we are throwing out more responsibility, but we are here to work with you to give you the tools you need and provide the training. But that talent strategy has to work with if that person has SME knowledge and has some sort of a data background, or at least, an inclination towards learning the technology. So that’s one aspect of talent on data governance.  


The other one, I would say, is around people who actually implement data governance platforms. For those like data governance engineers – somebody who has background in implementing, let’s say, Informatica, Colibra  or Alation – if you’re just an ETL developer, it’s a little different. It’s not exactly the same. So, if you know people who are doing data governance technology, it’s a specific skill set that they have. 


So, I would definitely go after that right skill set. I’ll give you an example in my instance. I got the right resource, and within literally one year, we were able to stand up six business units, mainly because of the fact that we already implemented that platform within that short amount of time. So, if you have the right resource who knows how all that works and is able to do that, I think that helps.  


And I always leverage a career parking approach. So, you have a lead, you have a senior, and you have a little bit of junior people on your team. And they pretty much have a career parking in that sense, as we want to promote them up the ladder. Also, it is a good mentoring opportunity with people who are the leads to mentor them and teach them. I think it’s a collaboration from talent strategy.   


Sue: That makes perfect sense. Sarang, thank you so much. It was such a pleasure to meet with you today, great insights. And to the rest of everyone, please visit for additional interviews. Thank you.  


Sarang: Thank you so much, Sue.  


This is the first part of a three-part interview with Sarang Bapat covering pressing data governance and strategy topics. Stay tuned for the release of the second part, coming soon, or learn more about our data governance thought leadership here 


Ready to get started on your data governance journey? Get in touch with Wavicle’s experts.