Enterprise deployment of agentic AI is moving beyond experimentation, but the hardest part is no longer building pilots; it is making them work as part of the wider business. In a webcast produced by MIT Technology Review in partnership with NiCE, Chandra argued that organisations should stop treating agents as isolated tools and instead design them as part of a connected operating model, with access to the right data, knowledge and systems if they are expected to make decisions and c...
Continue Reading This Article
Enjoy this article as well as all of our content, including reports, news, tips and more.
By registering or signing into your SRM Today account, you agree to SRM Today's Terms of Use and consent to the processing of your personal information as described in our Privacy Policy.
arry out tasks effectively.
That warning reflects a broader problem now affecting many companies. TechRadar has reported that agentic AI pilots often stall because firms try to layer autonomous systems on top of ageing infrastructure, fragmented data and weak governance. The result is a familiar pattern: promising proofs of concept that never progress into production because the surrounding environment is not ready for them.
Security and identity controls are becoming a particular pressure point. One TechRadar article said the rise of non-deterministic AI agents has exposed limits in traditional identity and access management, which was built for human users and static machine accounts rather than systems acting at machine speed. The piece argued that enterprises need a new identity model with tighter, action-specific permissions and stronger zero-trust controls if they want to expand use safely.
Governance is emerging as the central issue. Another TechRadar report said adoption is running ahead of oversight, with many organisations still failing to define clear boundaries, monitoring and approval steps for autonomous systems. The article suggested that without those safeguards, agents can create new silos, increase compliance risks and deliver little business value despite heavy investment.
That gap between enthusiasm and execution also shows up in survey data. According to ITPro, a Dynatrace survey found that half of agentic AI projects remain stuck at proof-of-concept stage, while security, privacy and compliance concerns remain major barriers to scale. Even so, spending is still rising, and most companies expect their AI budgets to grow further over the next year.
Chandra’s point is that the answer is not to expand indiscriminately, but to focus on high-value workflows and measurable results. The webcast argued that enterprises should build around a connected strategy that brings together use cases, workforce changes and outcomes, rather than “boiling the ocean”. The implication is that agentic AI will succeed less as a collection of clever pilots than as a disciplined redesign of how humans and machines work together.
Source: Noah Wire Services
