AI is often billed as a silver bullet, but a recent transformation at the US pension manager TIAA proves that the technology alone cannot fix deep-seated organisational problems. The firm, founded in 1918, spent years modernising its record-keeping infrastructure before it could reap any AI-driven gains.
Why AI cannot fix legacy problems
The analogy of a high-speed train on ageing rails captures the dilemma. Companies pour billions into cutting-edge models while still running them on outdated databases, siloed data stores and broken processes. The result is faster automation of dysfunction rather than genuine improvement.
Lessons from TIAA's transformation
When TIAA embarked on its digital overhaul, the first step was to retire legacy platforms that no longer supported the business. By cleaning data, consolidating systems and redesigning the way work flows, the firm reduced the time needed for plan sponsors to change investment options from weeks to days. Digital engagement across its millions of participants rose by 13 percent, a gain that came from infrastructure work, not a flashy AI demo.
Five steps for European enterprises
Industry analysts such as Gartner warn that only five percent of firms consider their data AI-ready, and that sixty percent of AI projects may be abandoned by 2026 because of data shortcomings. The following actions can help avoid that fate:
- Modernise the digital core before scaling AI agents. Identify which platforms are essential, retire the rest and rebuild a stable foundation.
- Make data readiness a prerequisite. A unified, governed data platform with quality pipelines is more valuable than simply adding more data.
- Redesign workflows, not just tasks. Map end-to-end processes first, then decide where AI adds value, applying an 80/20 lens to each role.
- Keep humans in the loop for high-trust interactions. TIAA's internal generative platform, GAIT, achieved 85 percent daily adoption among staff while reserving critical decisions for human agents supported by AI.
- Build for resilience and governance. Stay technology-agnostic, strengthen cyber defences and embed audit trails and human oversight, especially in regulated sectors.
These steps are not glamorous, but they are the foundation for sustainable AI impact. Companies that treat AI as a bolt-on risk spending years on pilots that never scale.
What comes next for European firms?
Enterprises across Europe that rewire their digital foundations now will be better positioned to compete in five years' time. The focus should shift from headline-grabbing demos to the less visible work of cleaning data, updating legacy systems and redesigning processes to suit AI's capabilities.