According to the "Value of AI Report 2026" conducted by SAP in collaboration with Oxford Economics, artificial intelligence (AI) is shifting rapidly from the experimental phase to real-world execution within enterprises. A survey of 2,600 business leaders across 13 countries shows that AI now supports nearly 30% of all organizational tasks, up from 25% last year.
While this increase reflects a widespread wave of AI adoption, many companies are still struggling to maximize the value of their costly investments. The boundary between successful adoption and failure depends heavily on each organization's systematic approach.
Background & Causes
Although AI investments are accelerating, most organizations still approach this technology in an ad-hoc and fragmented manner. According to the SAP report, more than half of surveyed businesses admit to investing in AI on a piecemeal basis, and only 17% have established a comprehensive strategic roadmap, though this figure has improved from 9% last year.
This fragmentation stems from board-level pressure on employees to quickly adopt AI without proper training, leading departments to implement standalone tools independently. This inadvertently creates isolated data silos and limits the ability to transform the entire enterprise's overall business processes.
Technical Analysis & Technology
To address performance challenges, the trend is shifting strongly toward "agentic AI" (AI agents) capable of multi-step planning and reasoning. According to SAP, the average general AI ROI rose from 16% to 21% this year, while the expected ROI for agentic AI jumped from 10% to 17%. A prime example is the beta testing of an accruals accounting agent, which reduces processing time from 12 hours to just 2-3 hours per month.
However, the biggest technical barrier for these agents remains data quality, with 73% of businesses admitting this is the primary reason AI falls short of expectations. Extracting data from core ERP systems often strips away critical business context, forcing providers like SAP to develop knowledge graph solutions to link millions of data fields without losing their original semantic meaning.
Expert Opinions & Insights
Sean Kask, Chief AI Strategy Officer at SAP, noted: "AI that lacks context, whether that's processes, data, or governance, at best creates activity without outcomes and at worst creates risk." Experts also warn about the emergence of "shadow agents" – self-operating AI tools accessing unauthorized data without the enterprise's knowledge.
Currently, only 12% of businesses report being fully prepared for AI governance, while 69% acknowledge that employees occasionally or frequently use unapproved AI tools in their daily work. This demands strict lifecycle management and access control solutions, similar to employee hiring and management processes.
Impact & Future
The shift toward the "Autonomous Enterprise" is reshaping global business operations, where humans and AI agents collaborate closely through natural language interfaces. However, the biggest challenge in the future lies not just in technical aspects but in human transformation. Around 80% of respondents agree that maximizing AI value requires more than technical upskilling alone.
For the Vietnamese market, where businesses are digitizing rapidly, lessons from this report highlight the importance of building clean data foundations and robust governance frameworks. Rather than rushing into expensive large language models, optimizing internal data structures and preparing governance capabilities will be key to the long-term success of AI initiatives.