VettaFi interviewed Microsoft Director Gaby Marano on data and AI.
VettaFi: Tell us about your background with data and AI.
Gaby Marano: I was lucky enough to begin my journey in data and AI 10 years ago when JPMorgan started its Big Data innovation lab. Then I ultimately became one of the firm’s first AI product managers and team leads. My mandate was to lead the research, design, development, and deployment of emerging technologies and to deliver them as integrated, profitable digital solutions for lines of business. Along the way, I ended up defining a lot of the AI product development frameworks used to start embracing probabilistic systems within the legacy tech.
Throughout that time, the data and AI tools in market were rapidly advancing — from public cloud services to computing infrastructure to machine learning to generative AI — so my role was to stay abreast of these changes, create focus on how they could impact the overall business strategy, put together complicated roadmaps with success metrics, and guide change management for adoption. I really loved my time there, but then mid-last year, I was offered an interesting opportunity to join Microsoft in the heart of their pivot to becoming an AI-led company. Now I serve as an industry advisor for financial services, partnering with investment firms and tech firms to solve the next big things for capital markets. Never a dull moment!
VettaFi: One of the challenges with businesses using data more effectively is that people have different degrees of awareness and comfort. Is there a way for more old-school industry leaders and the experience they bring to the table to incorporate data, or will there always be some philosophical tension?
Marano: I think finding the best solution will always require harmony between data perspectives and business perspectives. One example early in my career that cemented this was a late-night meeting where we had industry heads with decades of experience in the same room as skilled-but-fresh data scientists working together to forecast a market. We were all debating each other from very diverse perspectives and it was amazing to see how both schools of thought were so important to solving the problems at hand. I remember one executive who was quickly able to poke through the data analysis and explain where some underlying assumptions that only a seasoned professional would know were missing. And then the data scientists were able to pull out new trends emerging that the executives didn't expect to see. The collaboration was awesome and we were able to find clarity by the end.
In the age of AI, we all talk a lot about how much we love to innovate, but innovation is inherently uncomfortable. When working on initiatives that will transform and disrupt, trust becomes more important than ever between data teams and the rest of a firm. People need to believe the data is credible before they will agree to make decisions based on it or change their behaviors. It takes time to earn that trust, which requires proceeding at a pace that makes sense for the organization as a whole — even when some brilliant individuals can go faster. Before charging ahead, I always recommend focusing on ensuring all the functional stakeholders believe in the vision and commit to heading in that same direction.
The state of data in asset management
VettaFi: How would you describe the current state of data for asset managers?
Marano: Well, if you look at more consumer-focused companies, I think it is revealing that asset management is behind the curve in terms of what’s possible. My music app always knows what I want to listen to next. My maps app proactively assumes where I’m about to go and gives me the quickest and easiest route there. While asset managers get to enjoy this kind of personalized intelligence and acceleration in their home life, it starts to set a higher bar for their employee and client experiences. Alternative data has been a big priority for investment teams for a while now, but finding the signal in the noise is still difficult. It is exciting to see the distribution side in particular now starting to warm up and adopt GenAI, but overall I think we still have some more ground to cover across the space at large to truly see an impact.
VettaFi: What are some of the biggest mistakes companies make when they try to integrate data?
Marano: One big misconception I see is the idea that more data equals better outcomes. We see that debunked time and time again once organizations realize the need to have your different data sets talk to each other and the complexity in getting consistent answers to make decisions. What seems like a really intuitive business question traditionally can take days or even weeks of data management and cross-team dependencies to confidently pinpoint an answer.
Then after you overcome the underlying data problem, another place I see room to grow is in reporting. Having a report doesn't always mean you have an insight. People might be getting loaded with reports in their inbox every day that they marked as spam, or they have access to so many dashboards that they don’t use. Reports and dashboards are becoming easier to spin up in isolation but harder to integrate into a multidimensional, interconnected, holistic view into a business. There’s often a ton of strain placed on the data teams to predict and prepare every version of a question that might be asked of the data. Trying to build a one-stop shop for insight is a very tough spot to be in.
One more idea I see being challenged is that data is just your traditional view of numbers in rows and columns. That’s not the case; it is text too, especially in distribution. Our sales teams are out talking to folks all day long, and the content of those meetings might not appear to be data in the traditional sense, but there's so much we can do around different types of language data to extract that knowledge and share it across teams. It used to be a post-it note in a rolodex and now it's a call note in your CRM system, right? Well, now we’re moving to Teams meeting transcriptions with critical multimodal information that can automatically flow into sales and marketing processes, instead of requiring a ton of manual data management efforts to log what you learned.
Why AI could change everything for asset managers
VettaFi: Speaking of language data, let’s pivot to AI. What do you make of it as a tool?
Marano: I love it and I use it every day now. It's been really exciting now to see the push towards talking to massive amounts of data in a more conversational way. With the rise of Microsoft Copilot and other “ask me anything”-type tools, the barriers to entry for data analysis are coming down quickly and it’s easier than ever to learn almost anything. I still think we’re in very early innings though.
And back to data for a second — these GenAI chat tools are only as powerful and knowledgeable as the underlying data you connect to them. This whole craze really allowed so many more people to start thinking about data differently and come to the table to help design the future of AI-led businesses. How do we pass all of our enterprise intelligence back and forth in a momentous way where people are building off of the info from the person prior and workflows are moving seamlessly across teams in an organization? At Microsoft, we’re also focused on how agents can help fit into that puzzle where actions are taken autonomously on your behalf based on this fluid stream of endless input scenarios.
I think AI is here and it's here to stay. Agents are also right around the corner. The more you can try out some of these newer tools, share that feedback, and help build that efficient, well-executed relay race between data, sales, marketing — and even agents — the better.
VettaFi: Do asset managers have specific challenges and needs that data and AI could help solve? And what are the current obstacles?
Marano: The big theme I hear the most often is around how we use AI to allow employees and clients to self-serve as much as possible, so we can free up a lot of time to focus on the more important and exciting parts of the workday. In Microsoft’s recent Work Trend Index, we found that employees are interrupted about every two minutes between 9-5 and meetings outside typical work hours are up significantly year over year — creating a seemingly “infinite workday.”
If I think about why people were drawn to work as asset managers, it's typically not to spend hours perfecting the formatting and coloring of a deal doc or finding the dreaded #N/A across hundreds of Excel tabs or writing down all the action items during a long call, right? When you finish a great meeting and you feel jazzed about the ideas exchanged, you want to take action, you don't want to take notes. So how do we get the machines to do more of that mundane work and free you up for driving the next evolution of your business?
That leads into the theme of increasing employee productivity in asset management. Again, how do we automate away the “no joy” parts of the day and just double down on what's most important to drive more positive outcomes? And on top of all of that, how do we do this in a way that is compliant? Regulations vary globally. At Microsoft, we’re constantly talking to regulators about how to meet the requirements, exceed their requirements, and do everything in a way that puts privacy and security at the center.
So many mission-critical workflows still run primarily on Excel and email. There’s still a long way to go to unlock the full potential of this new technology and modernize asset management.
The importance of data optimization
VettaFi: How can firms better optimize their data so they can use it to effortlessly support sales and operations?
Marano: Getting to a place where you have a strong data foundation takes time. Is your data available? Is it timely? Is it in an analytical system? Do you have automated quality checks in place based on expected baselines? Are you aware of your blindspots? Are you aware of your conflicting and duplicative sources? Do you have people working every day to govern and derive insights from the data sitting in its raw form? How many different names do you have for the same client? How many different definitions do you have for the same concept? Is the data being used reactively, or are you using it to predict what’s going to happen next?
There’s a lot to consider, but once you get to that proactive state where data is embedded in the DNA of your workflows and you can start to predict with some compelling degree of accuracy where the business is headed, that’s when you know it’s working. At Microsoft, we call these “frontier firms” that are powered by on-demand intelligence and achieving the type of agility that generates value faster. You can feel everything start harmonizing. But it takes a ton of time and foundational investment to get there.
VettaFi: How do you get everyone in an organization on the same page about data?
Marano: It starts with genuine effort and intention to understand why people use data the way they do today and why they trust or don't trust it. Some people are what I would call a “super user.” They’re ahead of the curve, often ambitious people, willing to try out a new data tool or process, and give the data team feedback. They can be a champion for evolving your data and pushing the organization to embrace the power of this new technology. Those people can be amazing to work with. They'll give you the tough feedback that you need to hear to make your product better and can be the driving force between producing something good versus great.
But then, sometimes, you have others that are less intrigued and less engaged in the process. They're not willing to invest that kind of time. They will tell you to call them when it works and when it is guaranteed to have an impact, but they don’t necessarily want to be distracted with anything until then. I think, no matter where you might personally fall in this spectrum, it's really important for your data teams and product people to spend time understanding the various data personalities and incorporating all of this into the adoption strategy. We are in the midst of a massive evolution of organizational culture adapting to the new ways of doing things so it’s important we keep all of these different mindsets at the heart of what we build. Don’t trade speed for empathy.
