Signal Gartner expects over 40% of agentic AI projects to fail by 2027 on legacy constraints

Summary

Deloitte cites a Gartner forecast that more than 40% of agentic AI projects will fail by 2027. In the chapter on agents, the cause is traced to legacy systems unable to meet the execution demands of modern AI. Such systems lack real-time execution capability, modern interfaces, modular architectures and the secure identity management that genuine agentic integration needs. In the report's introduction, the same prediction is framed differently, with the problem located not in the technology but in organisations automating broken processes instead of redesigning operations. Deloitte summarises the lesson as redesigning rather than automating. As a remedy for data friction it proposes moving from traditional pipelines toward enterprise search and indexing built on knowledge graphs.