Why enterprise AI initiatives stall — and what CIOs can do about it

Gartner advises enterprises to avoid vendor lock-in for data, models, and workflows and to design AI stacks using open standards, open APIs, and modular architectures. The report warns that unmanaged generative AI technical debt will delay upgrades and raise maintenance costs for many organizations, and recommends maintaining an AI registry, enforcing model cards, implementing drift monitoring, and requiring vendor model-change notices to apply standard IT lifecycle management to AI projects.