1、RESPONSIBLE ENTERPRISE AI IN THE AGENTIC ERAExternal Document 2025 Infosys Limited Knowledge Institute3Responsible enterprise AI in the agentic eraContentsExecutive summary 4AI risks are widespread 6RAI enables growth 12Optimizing RAI 16Raising the bar 20Responsible agentic AI 24From compliance to g

2、rowth 28Appendix:Research approach 324Responsible enterprise AI in the agentic eraExternal Document 2025 Infosys Limited Knowledge InstituteArtificial intelligence(AI)is a game-changer,but it also brings risks.Responsible AI(RAI)the practice of developing and deploying AI systems ethically,safely,an

3、d transparently has emerged as a critical business capability in this context.The Infosys Knowledge Institutes survey and analysis of 1,500 senior executives involved in enterprise AI governance and execution in North America,western Europe,and Australia and New Zealand(ANZ)uncovers the scale of thi

4、s risk.This report also identifies many positives about how senior leaders view the need to deploy enterprise AI responsibly.Yet it also reveals large gaps in the ability of many enterprises to safely deliver AI implementations.Executive summary External Document 2025 Infosys Limited Knowledge Insti

5、tute5Responsible enterprise AI in the agentic eraAI risks are real and can be severeAn astonishing 95%of C-suite and director-level executives in our survey reported negative consequences of enterprise AI usage in their company in the past two years.These range from privacy violations,inaccurate pre

6、dictions,to bias or regulatory noncompliance,with almost all involving financial,reputational,or legal damages.More worrying is that almost three-quarters of companies cited damage that was at least considered“substantial,”with 39%claiming the damage was“severe”or“extremely severe.”RAI drives busine