Nearly two decades of experience provide ISG Research’s Matt Aslett with a unique perspective and expertise in data platforms, governance, management, and operations. Today, he has a front-row seat for the growing importance of data as organizations embrace artificial intelligence. But he chooses his words carefully when discussing the buzz surrounding AI.
“It’s a really interesting space because of the commitments being made and the money being spent, and the potential for change is obviously phenomenal,” said Aslett, ISG’s director of research for analytics and data. “But we do see many enterprises still trying to figure out where this makes sense and how they can achieve measurable performance, efficiency, or revenue gains. It’s not to say it won’t happen, but we’re in the very early stages.”
What’s evident to him, though, is how AI makes the category he monitors more relevant than ever for business success.
“Data is the foundation on which AI is built,” Aslett added.
With everything changing so rapidly, resources like the ISG Buyers Guides™ are essential in helping organizations make informed decisions. The UK-based analyst took time to discuss with us data trends he’s seeing and the recently published Buyers Guides for Data Integration and Master Data Management, both of which classified Boomi as an Exemplary Provider. We’ve condensed portions of our conversation for length and clarity.
Tell us about ISG Research.
Matt Aslett: We’re an industry research and advisory firm covering the IT services and software market. I work for ISG Software Research, which was formed in early 2024 when ISG acquired Ventana Research, which focused specifically on the software industry. We assess the software provider landscape in two main areas: One is enterprise applications like customer experience management, human capital management, and ERP. Then there’s IT, AI, and analytics. My focus is primarily on data and analytics within that group. I also collaborate with colleagues around AI because that’s an increasingly significant aspect of everything that we cover across the whole of ISG.
Why is data essential to any AI discussion?
Matt Aslett: Without quality data, you can’t train your models. You can’t perform the inference of AI models. You certainly can’t rely on the results if you’re not confident in the data being used within those models. This has been true of analytics since day one. But as enterprises have gone into the experimentation phase with AI, they’ve also recognized that there are challenges in terms of ensuring the data used is of the highest quality and that they can trust the results. Currently, many organizations are taking a step back to ensure they have best practices in place for data management, data governance, data integration, and data quality with a view to accelerating their AI initiatives.
Is data a root cause for stuck-in-pilot issues we see with AI?
Matt Aslett: In some cases, that’s being interpreted as why projects are failing. From our research, we don’t necessarily see that. IT projects fail all the time. Again, we’re in a phase where organizations are experimenting and innovating, and there’s naturally going to be a high level of projects that aren’t going to deliver the results you perhaps hoped for. But when organizations do take that step back, they’re making sure they get data management and governance capabilities in place.
Is AI forcing enterprises to reassess their data quality?
Matt Aslett: It depends on the use case. But AI increases the likelihood that you are combining data from multiple systems across an organization in order to automate. The problem of data silos is not new. But now AI is a board-level initiative and concern. Therefore, it raises the stakes placed on these projects. We know that with greater expectations comes greater risks. The last thing you want is for an organization to create an AI model that’s pumping out inaccurate information, or even worse, sensitive information from your customers, employees, or internal systems. It has really heightened the focus on the need to address some of these challenges, which have existed for many years, but perhaps didn’t have that higher level of criticality.
Does AI change the concept of trust in data?
Matt Aslett: To make business decisions, organizations need to be able to trust the data behind them. That was already a concern. We’ve seen a rejuvenation of data integrity initiatives as organizations try to address the trust issue. It goes beyond: “Is this high-quality data?” It’s something slightly more nuanced to say: “Do we trust this data?” I’ve had interesting conversations with clients around the difference. Increasingly, with AI, it’s not just about whether the data is fresh enough or complete enough. It’s whether enough people in the organization think that the data source is good. We see enterprises implementing initiatives to replicate with data products the sort of things that have become commonplace when people are buying retail products, like reviews and ratings. It goes to the confidence level. Enterprise-wide, not everyone needs to have a full understanding of all your data and where it comes from. But you have business users increasingly working on a self-service basis, being able to search for data within reports and dashboards. They do need to have a level of trust.
Help us understand the report’s methodology and what readers will learn.
Matt Aslett: We’re putting ourselves in the shoes of an IT buyer who is looking to evaluate potential providers for data management software. We go through a process of thinking about what someone in that role cares about. We’re looking not just at the software’s functionality, but also at all the platform capabilities that come with it. That means the reliability, manageability, adaptability, and usability. In terms of helping a potential buyer evaluate the value of that software we also look at the customer experience — what kind of ROI can they expect to get? How can they articulate the potential benefits to those further up the chain? Anyone involved in selecting data management software may not be the person who actually signs the check. So, there needs to be a process for communicating with others within the organization and making a case for whatever you select.
Are enterprises categorizing their challenges as data management or data integration? Or do they just say, “We have a data problem?”
Matt Aslett: When we come up with categories within the reports, we’re trying to reflect what we think is common in the industry. It is definitely interesting when you talk to enterprises, because you get different perspectives. Some of them definitely do think, “I’ve just got a data problem,” and they haven’t necessarily categorized it. Increasingly, we see providers addressing multiple data categories with a single platform, and Boomi is obviously a prime example. Broadly speaking, enterprises want to reduce the number of providers they’re working with. In some ways, consolidation makes evaluation more complicated because you have to pay close attention to what a platform contains when you’re comparing it to something else. But that’s something the Buyers Guides are designed to help organizations with. Enterprises get a sense of the market and identify which providers are in which categories. They can use it for their own assessments, or at least as a starting point. That is one of the things we want enterprises to get from the guides.
What do you enjoy about being an analyst?
Matt Aslett: It’s always changing. There’s always something new. There are established providers who are trying to evolve, do new things, and adapt to the market. You get new providers springing out of nowhere with interesting entrepreneurial ideas. There are always different ideas and approaches, new technologies, and something to learn. It’s a role I sort of stumbled into. I don’t know anybody who has ended up doing what they thought they were going to do. But it’s definitely kept me interested for the best part of 20 years.