AI Agents Transform IT into HR Department

The Rise of IT as Agent Managers

Agentic AI is fundamentally reshaping how enterprises approach automation, moving beyond simple helper applications to create what experts describe as a "parallel workforce". This shift means IT departments are evolving into something resembling HR departments, but instead of managing human employees, they're acquiring, onboarding, and supervising AI agents. At the recent Mobile World Congress, industry leaders highlighted how this transformation mirrors the microservices revolution that broke down monolithic applications into modular, independent units. **HPE**'s Bryan Thompson, VP for GreenLake product management, explained the architectural parallel: "Agentic AI is the next step in breaking apart and solving problems. There are opportunities to leverage these types of models and break them into almost like a microservice type of approach to tackle them, breaking them apart into specialised services." The implications extend far beyond technical architecture. As organisations adopt AI agents to transform their business, IT teams are discovering they need entirely new skill sets for managing autonomous digital workers.

Independent Execution Changes Everything

What distinguishes agentic AI from traditional enterprise software is its capacity for independent thought and action. **Deloitte**'s Abdi Goodarzi, head of Gen AI products, innovations and new businesses, emphasises this critical difference: "Until now, we've never had a technology that could ideate, or execute independently. Just think about that statement, and any other software package solution you've ever dealt with. None of them could independently execute any of it."

"Human capital management and agentic AI capital management are the same thing, but the difference is instead of an HR for humans, now you have an IT department that's acting as HR for all these agents." , Fred Devoir, Global Head of Solution Architecture for Telco, Nvidia

This new paradigm requires IT departments to develop capabilities traditionally associated with human resources: curating talent pools, establishing guardrails, providing training, and fine-tuning performance. **Nvidia**'s Fred Devoir articulates how this translates to enterprise workflows: "We take componentry and put it together into a RESTful architecture. Nvidia was able to optimise those with our microservices, and then bring together those microservices into blueprints to give a very quick time to value or time to first results." The cultural implications are profound. Goodarzi notes that unlike human employees, "Agents don't have emotions. How do you incorporate the emotions that will be part of the execution of the work?" This fundamental difference requires organisations to rethink collaboration models and how digital agents will transform the future of work.

By The Numbers

• Core HR headcount could fall by 30% or more as AI superagents automate hiring, training, and employee support
• 57% of HR professionals currently use AI tools for recruiting, with 35% applying it to resume screening
• CHROs project 327% growth in AI agent adoption by 2027 for autonomous payroll and onboarding tasks
• 52% of talent leaders plan to add autonomous AI agents to their teams in 2026
• Companies using human-AI augmentation achieve 2.5x higher revenue growth than traditional approaches

Solving Enterprise Data Challenges

Traditional enterprise systems create data silos that complicate AI implementation. However, agentic AI offers a novel approach by moving intelligence to where data already lives, rather than centralising everything. Devoir explains: "Instead of having to bring all your data to the AI, you're taking the AI to the data. When you make a service call, it actually asks all those data agents for a response and collates that data into a model." This distributed approach addresses several enterprise pain points:

• Eliminates costly data migration projects by working with existing systems
• Maintains data sovereignty and compliance requirements across jurisdictions
• Reduces latency by processing information closer to its source
• Allows incremental deployment without disrupting core business operations
• Enables specialised agents for different data types and business functions

The modular nature means enterprises can deploy agents incrementally, testing and refining their approach before full-scale implementation. This mirrors successful strategies outlined in Google's 5 AI agents to transform work by 2026, which emphasises gradual, strategic deployment.

Trust and Verification Challenges

The probabilistic nature of AI outputs creates new verification requirements for IT departments managing agent workforces. Unlike traditional transactional systems that produce deterministic results, AI agents generate probable answers based on training and context.

"Am I dealing with the right data? Am I dealing with the right results? Agentic AI is designed around probabilistic technologies. So you get the best probable answer because you have trained agents with a lot of knowledge on how to digest the data and make a decision." , Abdi Goodarzi, Head of Gen AI Products, Deloitte

This shift requires IT teams to develop new frameworks for quality assurance, audit trails, and performance monitoring. Organisations must balance automation efficiency with accountability, particularly in regulated industries where decision transparency is crucial.

Management Aspect	Traditional IT Systems	AI Agent Management
Deployment	Install and configure	Train, fine-tune, and onboard
Performance Monitoring	System metrics and uptime	Decision quality and learning progress
Updates	Patches and version releases	Continuous learning and retraining
Troubleshooting	Error logs and debugging	Behaviour analysis and bias detection
Compliance	Access controls and audit logs	Explainability and ethical guardrails

The maturation of agentic AI capabilities suggests 2025 will be a pivotal year for enterprise adoption. As Anthropic's Claude advances autonomous AI agents, organisations are gaining confidence in production deployments.

How do AI agents differ from traditional enterprise software?

Unlike conventional applications that follow predetermined workflows, AI agents can ideate, make decisions, and execute tasks independently. They adapt to new situations, learn from interactions, and collaborate with other agents to solve complex problems.

What skills do IT departments need for agent management?

IT teams require new competencies in AI training, bias detection, performance optimisation, and ethical governance. They also need understanding of probabilistic systems, continuous learning processes, and human-AI collaboration frameworks for effective agent oversight.

How can organisations ensure AI agent reliability?

Reliability requires robust testing frameworks, continuous monitoring of agent decisions, clear audit trails, and fallback procedures. Organisations should implement staged deployment, regular performance reviews, and human oversight for critical business processes.

What are the main challenges in agent deployment?

Key challenges include data integration across silos, establishing trust in probabilistic outputs, managing cultural change, ensuring regulatory compliance, and developing new governance frameworks for autonomous systems that operate across traditional departmental boundaries.

Will AI agents replace human employees entirely?

Rather than wholesale replacement, AI agents are creating hybrid workforces where humans and machines collaborate on complementary tasks. Humans provide emotional intelligence, creativity, and strategic oversight while agents handle routine, data-intensive, and analytical work.

As enterprises navigate this transition, the question isn't whether IT will become the HR for AI agents, but how quickly organisations can adapt their management practices to harness this new form of digital workforce. What challenges do you foresee in managing AI agents as enterprise employees? Drop your take in the comments below.