By Cynthia Lee, Vice President, APJ at Delinea
Artificial Intelligence (AI) has become pervasive across Indian businesses. Industries across banking, retail, healthcare, manufacturing, and IT services, have made AI an active part of their workflow and organizations. Businesses are no longer just deploying AI tools; they are now managing a non-human workforce.
Growing use of AI
AI is no longer just a tool limited to helping employees with completing their tasks. Now, AI has started taking on tasks independently across sectors. For instance, banks use AI to monitor transactions while digital platforms rely on AI systems to forecast demand and personalise customer interactions. Hospitals seek AI support for routine patient workflows. Technology firms are using AI to improve efficiency across different departments and teams. In some cases, AI is functioning more like human workers at machine speed.
As per a recent 2026 Identity Security Report, Indian organisations are feeling the pressure that comes with this shift. 40 percent of organisations in India say identity-related challenges have already delayed AI initiatives, the highest reported globally. The report also found that 79 percent of Indian organisations lack alternatives to standing access, while 63 percent believe identity discovery gaps in AI-related environments will continue to persist. These findings reveal a growing tension between the speed of AI adoption and the security controls required to manage it responsibly.
Barriers to progress
The challenge is many businesses still fail to understand that AI is not just a tool, but a workforce that can adapt to how an organisation functions. AI agents often interact with organisations’ important and confidential information, sometimes without oversight from IT and security teams.
Take the example of a financial services company deploying AI to process loan applications. The AI workforce may need access to customer records, credit data, internal approval systems, and compliance tools. If it operates under shared credentials or broad access permissions, the organization can lose track of who performed a decision and why. That is not just a technical problem but a business governance issue.
This is why organisations must start treating AI agents like digital employees. Every AI system should have its own identity, a clearly defined role and a responsible owner inside the business. Companies need to know why an AI agent exists, what systems and data it can access, and what decisions it is allowed to make.
Without that structure, businesses may struggle to answer basic questions about accountability.
In India, the urgency is even greater as the country moves toward becoming a global center for AI innovation. The pressure to move quickly is intense. But managing this new workforce requires a different approach to access. AI can work continuously and outside of normal business hours when humans may have more oversights of its actions. If these systems are overprivileged, a single mistake can be costly. Therefore, enterprises must use OTPs and short credentials to allow access only for a particular time period.
Offboarding – a hidden risk
Offboarding is another hidden risk. When employees leave a company, their credentials and information are usually deleted. But when an AI tool is no longer being used, the credentials are forgotten. These orphaned identities can become silent vulnerabilities inside the business. With AI adoption growing rapidly, businesses need to build a lifecycle management just like they build for humans.
Human oversight is critical to responsible adoption of AI. Every AI agent should still be linked to a person or team that remains accountable for its actions. Sensitive decisions should involve human review, and activity should be fully visible to security teams. Autonomy may reduce manual work, but it does not remove responsibility.
India has a growing workforce, and the future of work will not solely rely on people using AI. It will involve people working alongside AI agents that operate as part of the workforce itself. Once organisations master this shift, they will manage to innovate securely.