Why AI Training Fails and The Persona-Based Enablement That Works

Robin Sutara
The partnership with Databricks, as we think about how to make them successful, is not just with the technical platform. How do we also make sure that we’re investing in the other things that we know they need to be successful, like people, process? It’s not just about, hey, give me my data and let me go with it. It’s like, how can different people and teams work together to the common goal? Where I see organizations continue to struggle is not the expectations that they will have data available at their fingertips. It’s the, how do we change our way of doing business as a result of that insight. The way I have seen AI work effectively is to certainly be fast on experimentation, try things. You will never have your entire data platform. You will never have all of the data across your organization in a clean, pristine state. So I think if you are sitting back to wait, I think you will wait forever.

 

Saket Saurabh
Hey everyone, this is Saket and thanks for listening to another episode of Data Innovators and Builders. Today, I’m speaking with Robin Sutara, Field CDO at Databricks. Robin, thank you for chatting with us.

 

Robin Sutara
Yes, thank you, Saket. I’m really looking forward to the conversation today.

 

Saket Saurabh
Robin, you have an incredible background. I would love for you to share a little bit of that with our audience.

 

Robin Sutara
Yeah, so I’m really old. Hopefully it’s not too long, but as you mentioned, I’m Field CDO at Databricks. I have been with the company for about a little more than three years. Essentially what that means is that I get the opportunity to help some amazing customers as they think about how to be successful with their data and AI strategies, hopefully using the Databricks platform in some capacity, but really how do you think about it beyond the technology? So what I do is come in and support the partnership with Databricks as we think about how to make them successful, not just with the technical platform, but how do we also make sure that we’re investing in the other things that we know that they need to be successful, like people, process, operating models, change management, organizational design, sort of how do we help them drive that data and AI culture that they’re looking to accomplish as they think about innovation.

 

Prior to Databricks, I spent 23 years with Microsoft, where I did similar work helping Microsoft go through their digital and data transformation when Satya Nadella came in as CEO. I had the opportunity to do that in multiple roles, including Chief Data Officer of Microsoft UK, Chief Operating Officer of Azure Data Engineering, as well as managing a team of data analysts. So I’ve been at it from the ground level all the way up through execution to strategy.

 

I originally started my career at Microsoft in the late 90s supporting IE5 on Windows 3.1, so I did spend much of my career flipping between business and technology. And prior to Microsoft, I was in the U.S. Army where I did weapons and electrical systems on the Apache helicopters in Korea on the DMZ, so I love to tell people that my career has now come full circle from Apache to Apache.

 

Saket Saurabh
Very well said. What an incredible background, Robin. So why don’t we start with the topic of bringing AI to production, and you talked about big rocks in that topic. Let’s get right into that.

 

Robin Sutara
Yeah, so right now I spend most of my time on the road with organizations trying to figure out how they’re going to execute against this. I think right now is a fascinating time in that the last couple of years has really been a revolution in how organizations are thinking about not just AI, but they’re also going back to the fundamentals and foundations on how do they make sure that they have a good, strong data strategy in place as they think about executing against AI. I think we all know that you need good data to have good AI, and so for many organizations, they are thinking about, do we have the fundamentals in place to be able to actually use AI in production at scale?

 

What that means is, do we have data quality? Do we have good governance in place? Do we have agreement across the organization on the operating model of where data ownership and stewardship ends versus where the technology and tools are going to be able to support it? How do you balance the risk versus innovation? Everybody was super excited two years ago when ChatGPT came out, the pace of innovation is now going to be beyond reproach. For most organizations, there was almost this step back of how do we make sure that we’re minimizing the risk that represents for our organization?

 

So I think for most organizations that I work with, they are still looking at how do we break down the data silos? How do we make sure we have good data quality? How do we make sure that we’re getting alignment on the governance and the fundamental foundations that we need in place? Now, that isn’t to say that organizations are not doing AI. They absolutely are. I’m seeing lots of them tackling what are some of those where we have the fundamentals and the foundations ready and the data is ready for us to be able to drive AI. What are the business value outcomes that we can drive for the data that’s already in the process while we continue to modernize the rest of the estate?

 

I think the other component I talk to lots of organizations about is the people part. I think we talk about technology a lot, but how do we make sure that we’re thinking about the impact of the technology on the people, on the workforce, our customers, etc.? What does it mean to have an AI-enabled workforce? How do you think about what it means to have an AI-enabled executive versus an AI-enabled data scientist versus a plant floor worker at a car manufacturer? All of those are different personas. All of them have different capabilities and competencies. And what we expect from them when we think about AI enablement is very, very different.

 

So I see lots of organizations moving away from data and AI literacy that’s very broad in scope, trying to capture their entire workforce, and really starting to think about persona-based readiness and enablement. How do we make sure that we’re helping those within the organization translate what the technical capabilities mean when it comes to your job, to the processes that you do, to the functions that you serve, so that we can make sure that we’re enabling you and not replacing you.

 

Saket Saurabh
Yeah. And you work with some of the largest organizations out there. One thing I’ve heard from you is that while there is a myth and some studies saying AI is still in prototypes, you have actually seen people take AI into production and actually solve some real problems. Maybe let’s talk a little bit about that. What’s the real state? There are clearly challenges, data, people, but let’s start with the positives and the successes.

 

Robin Sutara
Yeah. I mean, I think for lots of organizations, while we love to talk about generative AI, what they did was actually use it as an opportunity to focus some of the efforts and activities around traditional AI and machine learning models. So if you think about particularly in the manufacturing space, or somewhere where you really could use deterministic models where you didn’t have the concern around the inaccuracies of the AI that you were rolling out into execution, we see phenomenal successes.

 

If any of your listeners want to check out the Databricks customer website, there are lots and lots of examples in manufacturing and financial services, in retail, healthcare. We just have some great customer examples of where they are actually in production using AI.

 

I think one of the big ones I would refer to would be JetBlue Airlines, because what I think is interesting about that use case is that they used the generative AI revolution, or evolution however you want to refer to it, to not only do traditional AI and machine learning, but they were also thinking about what is generative AI and what do agentic systems mean as they think about how they could actually innovate in an industry like the airlines, which has been relatively stagnant for so long.

 

I think they have done a phenomenal job of thinking about how do we do things like aggregation of data sets, like weather, aircraft maintenance, staffing, customer data, etc., to be able to do things like if a flight gets cancelled. They were thinking about how do we use agentic systems to automatically rebook customers so they don’t have to sit on hold or stand in line at the airport. Based on all of the things they know, where their aircraft are, which ones are down for maintenance, where they actually have pilots and crew that could staff the aircraft, where can they redirect customers and flyers to get on board. They are doing some phenomenal compound, complex agentic generative AI solutions in conjunction with traditional machine learning to drive incremental value for customers, and thus cost savings ultimately for the organization as well.

 

We also have some good ones in manufacturing, like Rolls-Royce, doing some phenomenal work with traditional AI around predictive maintenance on their engines.

 

So I think for lots and lots of organizations, I am seeing them roll out AI solutions, whether it’s what you would consider more traditional AI machine learning or whether it’s generative AI and agentic compound complex systems. Both are actually starting to really happen at scale, provided they’ve thought about where they have data ready that they can actually roll out these value outcomes. And the other thing is they really centered on what is the return on that investment. What is the business value that we’re looking to drive?

 

I think the organizations I see struggle the most are those that said, oh, we just need to talk to our data. Let’s set up a bunch of chatbots. We’ll figure it out later. They’re now taking a step back to say, how do we think about all of those investments? Are those things really driving incremental productivity, incremental revenue, cost savings, whatever it is that we thought we were going to drive? Are we really getting that out of some of the pilots? And if not, where do we de-invest so that we can reinvest in the ones that would have bigger impact?

 

Saket Saurabh
Yeah. And I think the way I have seen AI work effectively is to certainly be fast on experimentation, try things, and see what works. Of course, you’re all working towards making sure there’s business value being created. And one of the things you said, which is really good practical advice, is don’t wait for all your data to be in a perfect place. In fact, there’s no such thing. But find where things are in a good place and get started. Maybe give us a little bit of a double click into that. How should an organization think about finding those opportunities and not waiting for perfection?

 

Robin Sutara
I have to admit, I don’t know of any organization that is lacking in use cases coming from the business. I think for most organizations, we have thousands of use cases. Now, how do we look at it so that we can prioritize the ones that matter?

 

Whereas a few years ago, most organizations were focused just on experimentation, doing things, trying things out, figuring out what works, I think most organizations today are taking a step back to say part of that prioritization framework is that we have to have a clear return on the investment. We have to have a clear business impact before we’re going to start to test these. So there are lots of factors that can go into that prioritization, but some of the best frameworks I have seen are, do you really understand what the investment is? Do you understand what the return on the investment and the business impact is going to be?

 

And then as you think about the governance and the fundamentals that we talked about earlier, how are you as an organization balancing risk versus innovation? Because I think traditionally we would have seen highly regulated organizations like healthcare, financial services, public sector, government, etc., tended to be risk averse. And so you would have thought that they would have been the slower ones to adopt the new technology. And it’s actually not what we’re seeing. Financial services is probably one of the fastest moving industries that we see.

 

Jamie Dimon from JPMorgan Chase was recently on stage with Databricks at our Data and AI Summit in June. And he talks about how critical AI has become to their day-to-day function at the C level and within the organization, to the point where he actually has AI representation as part of his executive staff to make sure that they are thinking about how are they rethinking some of those legacy traditional processes and where can they leverage AI.

 

But I think one of the things that JPMorgan did really well was think about where is that balance, what are the risk factors that really matter to the organization where they needed to have absolute controls in place versus where are there opportunities to innovate? Because they were either doing things in a legacy way or the risk that it represents is less than what they thought it was.

 

So I think organizations have said absolutely no experimentation with AI, which is where they were a couple of years ago. And then there were other organizations that said do whatever you want. And I think for most organizations today, they’re trying to figure out what is that balance.

 

We have a great framework available on the Databricks site. It’s called the Databricks AI Security Framework. It was created by our field CISO and his team. It’s a phenomenal framework that lists out 62 risk factors. Before you panic, I promise not all 62 apply to your organization. But it does give you what are the risk factors we see across all industries. NIST and some others are using this as sort of a standard framework as well. So it’s a great place to start to look at which of these risk factors you really care about within your organization and what are the controls for each of those risks that you can put in place to make sure that you’re not exposing yourselves, your organizations, your patients, your clients. How do you make sure that you’re allowing the pace of innovation and yet protecting the things that matter to you as a company or as an organization?

 

Saket Saurabh
Yeah, absolutely. And I love that there’s a framework because I’ve seen people really worry about governance. And when governance becomes the driver of everything, it slows down the innovation. You’ve got to find the right balance.

 

Robin Sutara
Yeah. And for most organizations, I think it introduces incremental complexity because what will happen is you’ll have the business that want to innovate faster. Whether it’s legal or compliance or security or whoever it is, they feel like those teams are precluding them from innovating at the pace they want to. Often I see that’s where these shadow platforms, shadow IT gets stood up because the business feels like they have a sufficient value proposition on the use cases they want to roll out.

 

So this is where I think when you can bring IT and legal and compliance and government affairs and security and the business all together to say, what are the risk factors that matter to us? How do we now create a prioritization framework for all those use cases based on the risks that we are willing to take? Which are the use cases that you can actually proceed forward with because they’re not a risk to the organization?

 

Sometimes giving that empowerment to the business, we talk about democratization of data and AI a lot. To me, that means empowering the business to be able to leverage the technology on their own without having to go back to a centralized team, so that poor team doesn’t become overwhelmed and inundated with a backlog they just can’t keep up with.

 

Saket Saurabh

Yeah, and I often think of this as democratization more as like collaboration. It’s not just about, hey, give me my data and let me go with it. It’s like, how can different people and teams work together to the common goal? And I think that really brings us to one of the topics that I feel is really timely, which is AI literacy across the teams. You mentioned legal, business, engineering, all of those different teams that are trying to work together. How should we think about AI literacy and awareness across all of these?

 

Robin Sutara
Yeah, I think this is where I go back to the commentary around every persona across your organization is going to have a different need when it comes to literacy. I think years ago when everybody approached it from the AI 101 or the data 101, we really did one of two things. We were either too high and it was just a check-the-box activity, or other organizations immediately went super deep in the technology, expecting that everyone across the organization was ultimately going to be a data analyst or a data scientist in some capacity.

 

And I am actually finding for most organizations, it is not those personas that we should be focused on only as we think about data literacy. Yes, we need to help our data personas as the technology continues to move at the pace it does. But as you said, for democratization, it will vary depending on who the persona is. You could have an HR business partner or a finance business partner or a plant floor manufacturer or a store manager in retail, but all of them will have a different necessity for what it means for them to be literate. What does it mean for them to be AI enabled?

 

So as you think about your organization, my recommendation would be to look across the different domains and try to define as many of those personas as possible. There might be overlap. For example, your HR business partner might be very similar to another business program manager. Those personas might be very similar.

 

If you think about starting to pull together your entire organization into these personas, not by people or organization or roles, but really by personas, to say, what do we think is required for them if they were AI enabled? Does that mean they’re going to work in Perplexity and Glean and write Power BI or Tableau or whatever the interfaces and tools are today? Does it mean we’re going to leave them in that, but we expect them to do their processes or their job or their role differently? Then that’s what the enablement for those personas should look like.

 

It shouldn’t be, let’s put you into a Databricks notebook and make sure that we keep that code in Python. And so I think those that are doing it really well are starting to look across the organization to start to define those cohorts of personas, build out the enablement to say, if we were to say we were successful as a company or as an organization and they are AI enabled, what do we think those outcomes look like? How do we actually start to build the enablement that allows us to measure the maturity and competencies today versus where we aspirationally want them to be as a democratized AI capability? I think anybody that’s looking to just pull this off the shelf someplace without thinking about what are the capabilities and competencies within their own organization will continue to struggle.

 

Saket Saurabh
Yeah. I’m curious how things have changed from your perspective. We were in a very analytics-first world where for a lot of functions, understanding data might have been a bit harder, but now in the AI world with a chat interface, I think data is at everybody’s fingertips. Has that helped in terms of driving that awareness of data, that literacy of data and AI? Have people developed more ease of understanding across different functions where prior maybe they were not thinking about data as much?

 

Robin Sutara
I think more people are aware of it, because it’s almost become this expectation that you will have that information at your fingertips. I mean, I don’t think there’s probably a single listener that right now hasn’t thought about using ChatGPT to do something. I use ChatGPT almost on a daily basis, but it’s things like I want to remodel my kitchen, like if I change my kitchen cabinets, am I going to like the color? It has now become a given that I don’t have to go into Adobe and build it out and do all this graphics work. I can just put it into ChatGPT, snap some pictures, they change the color and let’s go.

 

I think that same thing has translated into the enterprise. For most organizations, there is this expectation on how easy it is to find information, to process information, to make a decision. Where I see organizations continue to struggle is not the expectations that they will have data available at their fingertips in natural language or whatever interface they want. It’s the, how do we change our way of doing business as a result of that insight?

 

Same problem we had, I think, in the data analytics days where data was accessible, even if you went into whatever dashboard or whichever of the 700 dashboards you had across the organization. It wasn’t that data wasn’t available. What the issue was is that we didn’t help the organization figure out how to manage the transformation and the change management that it would take to do their job differently as a result of the data. How was it not just backward-looking reporting out of what happened in the business, but how was it actually now driving decisions in a different way or changing a process or impacting end-to-end supply chain or last mile delivery? Avoiding stock outs at a retailer, like there’s all sorts of value proposition, but it was always backward looking as opposed to being predictive and preventative.

 

And I think we see the same thing with AI. Yes, there’s this expectation that I will have this data and AI at my fingertips in whatever interface I want. But now we have to help the organization think about how does that change how you do your function or your role? How does it make you more productive? How does it allow you to solve that backlog of business problems that you’ve been trying to solve but never had time for because you were trying to rationalize 27 Excel spreadsheets to figure out what to order for your store for tomorrow?

 

And that’s what we need to help with. When I think about persona enablement, it is that translation, to help those personas figure out, now that you have the power of all of this technology and all of this data and all of this AI at your fingertips, how do we help you figure out what does that mean for you in your day-to-day function? What are the things that you no longer have to do that you didn’t like to do in the first place?

 

My guidance for those on the data science or data analytics team is to go sit with the domain. You need to sit with those business users, because all too often I find these centralized teams sort of build these solutions thinking that they understand what it looks like to be a day in the life of a line worker at an electric company, but you don’t until you sit in the truck with them and drive from station to station and try to do manual verification of transformers. You don’t understand what their life looks like. So this is really where that business connectivity with technical capabilities represents so much opportunity for IT teams to work together with the business.

 

Saket Saurabh
Yeah, I think it sort of asks for almost everybody to become a little bit of a product manager, to spend time with your customer, understand what they’re trying to do, and really bring that back into how you’re solving that.

 

One of the things I’ve heard from you is that you see different patterns across the organizations you work with. Some will jump right into it, some will do a lot of planning to get into that. What elements of organizational culture play into that and are there specific approaches you would recommend depending on what that organizational culture is?

 

Robin Sutara
Yeah, I think I have yet to see any organization that can just jump in, unless maybe you’re an organization of three people and you figure out what that means over coffee. For most organizations there’s some level of complexity, some level of legacy, some level of history or change. Change is always the hardest part. Most people, I think we’re all guilty of not sticking with a new year’s resolution past February because change is hard when you have to change those behaviors.

 

So I think some of the best practices I see are the education and enablement that we talked about, like how do we make sure that we’re not creating the same expectation for everybody across the organization? And then once we know what that enablement is, how do we make sure that we’re giving them the things that allow people to feel more comfortable with the technology? Because as fast as it’s changing, we can’t constantly change the tools they’re using or the way they’re interacting.

 

How do we make sure that we’re maintaining where we can, that we’re allowing the same interface? For example, there’s a great partner called Sigma that has an Excel front end with a Databricks back end, because people want access to that data in Excel. We don’t want it stored localized where people don’t have the value proposition that data represents for the rest of the organization. So start to think about, when you think about those personas, where are the tools? Where are they comfortable today? Where can you minimize the change while still modernizing and getting the data in a way that allows you to actually drive AI use cases on the other side?

 

I think the other thing is thinking about a safe space to experiment. I don’t know of any organization where every data and AI use case they’ve rolled out has been successful from the very beginning. So for many organizations where they’ve created a safe space where leadership doesn’t have an expectation of a 100% success rate for every use case or proof of concept or pilot, what is an acceptable level of failure and how do you take the learnings from that failure to figure out what you’ll do in the next use case?

 

How would you roll that out across the organization differently to help the organization adopt the capability or the technology? Where is there the absolute we need everybody in the organization to use or leverage some capability or competency that you’re building? How do you think about the processes and the frameworks that exist within the organization that you need to change to reinforce that behavior? How do you think about things like performance management across the organization? If you want the entire organization to change and use data and AI differently, have you changed their KPIs and the expectations on how they perform in their role?

 

So I think the organizations I see doing it really well have created that safe space for experimentation. They’ve thought about the impact that the actual pilots they’re going to roll out at production scale will have within the organization, whether it’s a capability or a competency that you’re going to build in that persona, or things like process, performance, metrics, measure, anything that you need to change to help reinforce that behavior.

 

But I think that goes back to traditional change management that’s been around for decades. So I think oftentimes we think it’s a magic bullet that’s going to solve all these problems, but taking that step back to say, how do you create awareness or desire or knowledge and reward the right behaviors and continue to reinforce the capabilities and competencies that you view as successful for the organization?

 

My recommendation for that is figure out what has worked within your organization as you think about any big change that you have rolled out, whether it’s a digital transformation, whether it’s a technology transformation, whether it’s a new executive leader. There are lots and lots of changes that organizations go through all the time. Look and see which ones have been the most successful and what were the things that within your organization, within your culture, resonated to help you roll out that transformation in a way that was successful. And how do you replicate some of that as you think about data and AI transformation going forward?

 

Saket Saurabh
Yeah, I think the guide to future success might lie in prior areas where you have seen that success and have seen what has worked in your organization. And talking of that, digital transformation and data transformation have been big topics. Where are we on that? Has that become easier? Are there still major roadblocks for people working through that? Or is that just a continuous process that we have gotten used to as we are building?

 

Robin Sutara
I think there are probably organizations out there that would say they absolutely haven’t figured it all out yet. I think about the pace of innovation that we see on the technology side. If you think about it, in years past when I first started at Microsoft, the mission of the company was like a PC on every desktop. I remember one of the things that Satya Nadella did was change the mission because by the time he came in, everybody had a PC, or PCs were obsolete. We had already moved off to mobile devices and there were other form factors being implemented when it came to technology.

 

So if you think about things like the Internet, like PCs, even mobile devices, took years or decades to actually become norm and mainstream. The interesting part is that I think the pandemic was probably the first thing that really happened to us where we saw an instant need to transform as an organization in digital and data capabilities. Lots of organizations had to use that as a way to figure out how are you going to change your operating models, your business models, how are you going to transform the way that you interacted with customers or clients or patients? How do you interact with your employees differently or how do you give them capabilities and competencies? That was interesting in that, unlike other transformations in the digital and data worlds that we had seen before, where you had years to take the organization through it, the pandemic was a great example of the resilience of your organization to do almost instantaneous transformation and change.

 

I think I view generative AI and ChatGPT very similarly in that, unlike most technologies that took a while to be adopted, it was one that was almost instantaneously available to everyone. And there was an almost instantaneous translation that lots of people within the organization saw as an opportunity to transform the way they did their job and the way that they could bring it into the enterprise. So unlike a PC where for most people there was very little interaction between work and home, this was something that was really transforming the way organizations worked.

 

So I think if organizations can think about, how did you respond during the pandemic? How did you drive your organization through that change? Yes, it won’t be as painful. We’re not going to put you in quarantine or make you wear masks all the time. But how do we think about some of the things that really worked as we were responding in real time to what was happening? I would view this transformation as very similar. We’re responding in real time to the capabilities of the technology. And that doesn’t mean that we will automatically change everything about the organization, but it is a great opportunity to take a step back and say, what are the things that we can change or should change as a result of the technical capabilities that exist today that didn’t exist 10 years ago?

 

Saket Saurabh
Yeah. I mean, you work with many organizations of different sizes and different industries, and you’re coming in as a CDO. Many of them may or may not have a CDO, but you’re coming in as an outside expert. Do you have some quick recipes for people, like early success in 30 days or 60 days that you might want to suggest? Like you’ve given quite a few of those, but is there one or two things you would say, do this and you’ll see some success?

 

Robin Sutara
Yeah, I think it’s what we talked about earlier. How do you focus on that quick thing? You will never have your entire data platform. You will never have all of the data across your organization in a clean, pristine state. And so I think if you are sitting back to wait until you have everything aligned and everything ready and everyone in the organization bought into the change and the transformation, I think you will wait forever.

 

So absolutely my recommendation for organizations of any size that do this really well is to be focused on, of all these use cases received from the organization or from the business, which are the ones that you can actually deliver against today? Which ones have the data ready, the right governance in place, the right operating model and agreement on how it’s going to work, the right enablement plan in place for the actual users of the solution? Where are some of those quick wins that you can really deliver?

 

And I will go back to the commentary around the maturity of the organization. I have seen organizations take the most mature parts of their business and focus on those use cases because the lift to get them enabled and actually change and transform is relatively light, especially if it’s a complex use case. If you’re talking about a simple use case and you have a less mature part of the business that you’re trying to take on the journey, but the tool or the interface is natural language or something that makes it easier, and you want to use that as a quick win to start to get the rest of the organization bought into why the transformation, that works too.

 

So again, I think it’s sort of a case-by-case basis. You need to know both the maturity of your organization by domain and by persona, and the aspiration when you look across the organization, before you decide which use case is going to be the most impactful. Otherwise, it’s really easy to just tackle one use case after the other, but it’s not always the right ones for the business. So how do you start to think about where you are today, where you want to be in the future, and then how do you start to prioritize some of those use cases based on the fundamentals and foundations that are in place that you could actually deliver?

 

Saket Saurabh
Yeah, and you definitely touched earlier on the importance of looking at the ROI and understanding and measuring that. And one of the great things you noted was in organizations like JPMorgan, where it’s coming from the CEO, and I think that has a huge impact in driving change across that organization and bringing momentum to that.

 

Before we wrap up, I was wondering if you might have some pointers for folks in terms of places they should look at for learning and getting better at understanding the organizational aspects of data and being able to move faster with change. You shared a couple of case studies, but anything else you want to point us to?

 

Robin Sutara
Yeah, I think there are always lessons learned from the past. If you think about traditional change management, there are some great courses out there. You can do it as a certification program, lots of places. There are lots of great programs around just traditional change management. And I think for anyone looking to drive that change, understanding those fundamentals of how to drive change or transformation is critical. There are some great Harvard Business Review case studies that exist on successes and failures. I think Microsoft has one out there that’s a decent read.

 

I also find things like the Acquired podcast really helpful. It talks about acquisitions that have happened across the ecosystem and the change that was required. Mergers and acquisitions are huge transformations for organizations. So I’m always fascinated to listen to some of those stories on which acquisitions went well or didn’t go well, and can I translate them back into an organizational change management sort of failure or success?

 

So think about where you can learn from the past, where we have tried or seen organizations try to change and transform over time and what has worked for them. Because sometimes it gives you an idea, especially if you can compare to an organization that’s similar to yours in size, scope, industry, clients, or geography. Can you look at what you would view as successful acquisitions that have happened in your space and do some comparison on what worked or what didn’t, as you think about building out that muscle and capability within your own organization?

 

Saket Saurabh
Well, thank you. That’s a great piece of advice. And finally, you’ve been very active in giving back to the community as a mentor to data and AI professionals. In particular, you have done a lot of work empowering women leaders in data. Anything you want to share with the audience and how they can be thinking more actively about giving back and contributing?

 

Robin Sutara
Yeah, I would say data and AI is a small world. So anywhere where you can help build the community, what I really enjoy about the participation in the community is how do we make it better for the next generation of talent coming in? I definitely want them to learn on their own and make their own mistakes and not just have to repeat all the many mistakes that I have made throughout my career.

 

So I think there are two benefits. One is how do you help the next generation grow and develop and go even further than we’ve been able to do in our lifetimes? And then it’s a super small world. So anywhere where you can actually drive value, share best practices, talk to somebody else who’s going through the same pains that you are, figure out the communities that exist across your organization.

 

I highly recommend finding a community that includes people from outside your company. Oftentimes we get so caught up in work and the day-to-day job, and everybody has that same viewpoint because you’re all within the same organization. So anywhere where you can connect with an outside community, whether it’s a Databricks user group within your city, or Women in Data, or Women in Tech, or whatever it might be, find one or two communities where you really can connect, because you will find that you will stay connected with those people throughout your career as you all move between organizations and deliver more and more successes. Hopefully we can all lift each other up across the data and AI community that way.

 

Saket Saurabh
Awesome. That’s a great message to wrap this up with. It’s been an incredible conversation chatting with you and learning from you. And before we wrap up, if folks want to follow along with your work, what’s the best place for them to do that?

 

Robin Sutara
Yeah, LinkedIn is probably the best place. Feel free to reach out. We’d love to hear from you. And if there’s something that resonated or if you want to have a conversation, if you don’t agree with me, I would love to have a discussion as well.

 

Saket Saurabh
Awesome. Yeah, I love controversial conversations. We learn a lot when we do that actually. So thank you so much for joining us today, Robin. Thank you.