Index Insider: Why AI Needs a Data Tower, and What It Means for the Sector | ISG

Hello. This is Stanton Jones and Alex Bakker with what’s important in the IT and business services industry this week.

If someone forwarded you this briefing, consider subscribing here.

What You Need to Know

For enterprises that want to use AI to achieve business goals, managing and modernizing their data is a critical requirement. And many leaders say their current efforts are underperforming. This is why enterprises should consider managing – and sourcing – their data initiatives as a distinct “data tower.”

Data Watch

Background

AI’s rapid evolution—from predictive to generative to agentic—has transformed how enterprises need to handle data. However, unlike predictive AI, which depends on carefully curated data models, generative AI’s reliance on massive unstructured datasets is putting traditional data management frameworks and activities under tremendous pressure.

AI Requires New Ways of Working with Data

We’re looking at some of these activities in today’s Data Watch: on the X axis is the level of investment in data management activities; on the Y axis is the level of performance of these activities against expectations.

• Activities in green: Areas like data security, compliance and backup/disaster recovery are performing well. Notably, all these activities are reacting to external factors, like third-party best practices and regulatory requirements.
• Activities in blue: These are some of the core data management and modernization activities required for the effective application of AI to business problems. These include managing access, ensuring data is accurate and understanding the business meaning in the data.
• Outcomes in red: Realized AI productivity gains is the ultimate goal of many AI programs, but without mature data management in place, most business goals – whether it’s improving productivity, reducing costs or growing revenue – are unlikely to perform well. So what we’re seeing here is lots of investment in areas that ultimately won’t improve business results.

What Does this Mean for the Sector?

We see significant opportunity for service providers to help enterprises with these data management and modernization activities – what we’re collectively calling the “data tower.” In fact, we’re already seeing a number of sourcing deals in the market structured this way. And over the next 24 months, 60% of enterprises plan on adding service providers to support these data tower activities.