One of the report’s key takeaways concerns security: 93% of CIOs and CTOs believe that running AI on their organization’s legacy systems without first modernizing them will eventually trigger an enterprise-wide security crisis.
The leaders we surveyed aren’t just worried that a security crisis might, or could, happen. They believe it will happen if they don’t modernize.
Legacy systems can feel overwhelming, especially in large enterprises, where they thread through core applications, data environments, integration layers, and other central infrastructure. Why transform something that’s working?
But as our survey shows, legacy systems present numerous challenges in the medium and long term. And it’s important to recognize that modernization shouldn’t aim to transform an organization’s entire tech estate at once. Instead, it needs to be intentional and methodical.
Leaders must determine where AI can add real value to their business, and then investigate the systems involved. What data would the AI need to access? What dependencies need to be accounted for? The resulting operational map enables systematic modernization.
That’s even more important when it comes to cybersecurity. As AI becomes increasingly autonomous, the architecture and governance surrounding it will only grow in urgency. Protecting the data – and the trust – of enterprises and their stakeholders comes down to baking security directly into the foundation on which AI is deployed. That means simplifying integrations, improving data controls, and building for the next phase of innovation from square one.
We’re already seeing what this methodical approach looks like in practice. For the financial services arm of a global automotive manufacturer, GFT built a secure cloud landing zone as the foundation for its transformation, enabling the organization to safely migrate its core banking platform and then move additional workloads according to its business and technology priorities.
The example points to a lesson: Modernization does not have to mean transforming everything at once. It means understanding where legacy complexity stands between the business and the value AI can deliver, then modernizing deliberately around those priorities, in a secure and efficient way.
The AI implementation gap is not simply an AI problem. Increasingly, it is a question of enterprise readiness. The organizations that close the gap will be those that treat modernization, security, data, and governance not as separate technology initiatives, but as the foundation for turning AI ambition into sustainable business value.
