Transition toward agentic AI in public and enterprise use

The landscape of artificial intelligence is transitioning toward agentic systems capable of executing end-to-end workflows and specialized applications. In the public sector, government agencies are prioritizing operational efficiency through digital transformation, though progress is often limited by budget constraints and siloed decision-making.

In the enterprise sector, the focus is shifting from simple prompt engineering toward robust agentic platforms and advanced Retrieval-Augmented Generation (RAG) systems. The effectiveness of these systems is increasingly tied to the quality of document pipelines; as companies integrate unstructured data like PDFs and scans, they face the challenge of preserving context, such as table relationships and document hierarchies. While initial interest was driven by productivity tools, organizations now face pressures regarding rising token costs and the need for meaningful ROI.

Industrial organizations are specifically turning to agentic, AI-powered ‘digital workers’ to address workforce capacity shortages caused by the retirement of experienced personnel. Research indicates that industrial employees spend roughly 41% of their time on repetitive manual tasks, and 77% of decision-makers have delayed strategic initiatives due to lack of team capacity. Although investment is accelerating—with two organizations preparing to invest in digital workers for every one that is not—a significant trust gap persists. Only 5.7% of decision-makers trust AI to act fully autonomously, reinforcing the continued necessity for human-in-the-loop models where humans provide essential judgment.