McKinsey refers to technical debt as the silent killer of technology modernization, accounting
for about 40% of IT balance sheets and adding an extra 10% to 20% on top of project costs. In many cases, this leaves limited room for strategic investments.
Legacy systems weren’t built to work together. They don’t talk to each other or share data — causing friction across processes, people, and technology. At the same time, proprietary applications and custom development often prevent IT teams from taking advantage of software upgrades, among other things.
AI tops the priority list of virtually every C-suite, but technical debt gets in the way. It’s not the algorithms or models themselves but the decades of deferred decisions in enterprises’ tech stacks. This may help explain in part why, according to a 2025 MIT report, 95% of organizations are getting zero return on their AI investments. AI technology is ready, but the underlying infrastructure isn’t.
Mobilize leadership around strategic outcomes
As CIOs become more business-oriented versus technology-focused, they must galvanize organizational leadership around reducing technical debt. This includes articulating the business value in doing so — helping others in the C-suite understand that tech debt remediation is a critical step toward getting reliable, sustained ROI from AI solutions and toward the enterprise’s ability to more quickly adapt to future waves of change.
Tech debt can also help shape an organization’s initial AI strategy, highlighting areas and processes that are already ready for AI to go to work, plus areas that will need attention before AI is deployed. This can preempt issues with scaling solutions in areas that aren’t organized or mature enough to handle AI. Patrick Thompson, Global SVP, Customer Transformation, at Celonis, offers this advice. “If you’re going to invest in AI, it has to have a return. You need to figure out the organization’s pain points and where AI can really get the most value,” he says.
Conversely, AI can also help IT teams in the process of tackling debt remediation and help them prevent it from compounding as companies adapt to the exponentially increasing pace of change. Per Thompson, teams should ask themselves, “Where do humans need the most help, and how can you create a 5x multiple of people by using AI? How do you make them 5x or 10x more efficient or enable them to scale as the company is growing? It’s about finding those use cases within the business.”
Context is everything: Why artificial intelligence needs process intelligence
Gaining the most value from enterprise AI first demands operational context. It’s the semantic glue that binds data to meaning and relationship, enabling users and systems alike to derive true insight.
But contextualizing CRM, ERP, human capital management (HCM), supply chain management (SCM), and other systems of record is difficult. System data is not engineered to show how the systems all work together. Thompson says CIOs must contextualize systems of record before running AI. “Make sure you understand the power of your data and your systems of record. That is what will enable your AI strategy.”
This is where the Celonis Platform, powered by the Celonis Context Model, comes in. Simply put, per Thomspon, the Platform acts as a missing context layer for AI, helping solutions and people work together to continuously optimize operations — while taking into account each business’s unique way of working.
“If you’re taking 20 steps between an end-to-end customer delivery or order on time, Celonis can give you ‘process intelligence’ — it can show you those steps in your process, but also show you only really need 10 of those steps,” he says. “Why are you doing these other 10? And that takes cost out, that creates agility, and that creates productivity, especially when AI agents can help you actually make these changes.”
Modernization programs address tech debt, deliver value, and scale AI
Without the right business context — the context enterprises can get from the Celonis Context Model — AI projects will continue to suffer from technical debt. Technical debt is one of the major causes of disconnected, obscured tech stacks. From LLMs to agents, AI can’t successfully function at scale if it doesn’t have a reliable baseline of information about how systems and processes in the business are being used. So, without a real-time way to understand how things are happening at your business — who’s using what to do what — AI solutions will be much more error-prone and much less able to reliably work as agents.
Celonis offers a baseline of where such debt causes real harm and serves as a source of truth for effective AI deployment. Insights into operational roadblocks that slow people down, inhibit innovation, and limit modernization projects can help identify where to address technical debt – and which areas will yield the most immediate ROI.
Solutions such Celonis enable IT leaders across the many different stages of modernization to rectify technical debt throughout. This platform supports broader modernization strategy development and focuses on creating business value, not just paying down debt. The platform can help IT leaders manage and execute transformation programs across the enterprise more transparently and effectively while ensuring that changes remain in place.
These different aspects of a modernization program, when done intelligently, can work in service of the full-tilt AI transformation that businesses require for long-term success. With fragmented data turned into a shared contextual view, IT and the business can align on what’s working, what’s broken, and where to begin tackling technical debt.
The bottom line
IT leaders who want to turn AI hype into ROI must address technical debt. By design, the Celonis Platform is the foundation for identifying and tackling such debt with the most return on value. The operational context provided by the Platform is also fundamental to implementing AI at scale.
To learn more, click here.