In human resources departments, the past few years have been a race to the cloud, with teams migrating to centralized human capital management (HCM) solutions faster than ever. Yet for many, the move to modernize has happened in name only. They’ve shifted platforms without realizing the full agility, efficiency, or user confidence those systems promise.
HR leaders have turned to artificial intelligence (AI) to close that gap. Tools like generative assistants, predictive analytics, and workflow automations have helped recruiters screen candidates faster, supported HR in writing job descriptions and policy updates, and enabled finance and payroll teams to forecast workforce costs with greater accuracy. These advances have driven the first wave of cloud HCM transformation, making work faster, smarter, and more connected.
But for most organizations, these tools still operate on the periphery of the employee experience. They assist, but they don’t act. That’s beginning to change with the rise of agentic AI — systems designed not just to respond, but to decide, adapt, and execute across complex enterprise workflows. It marks a shift from assistant to autonomous co-worker, with the potential to turn cloud HCM from a system upgrade into a new way of managing work itself.
Still, Gartner predicts that more than 40% of agentic AI projects may be canceled by 2027 due to unclear value and risk. Success in integrating this digital workforce will depend not on mere innovation, but on governance, measurable outcomes, and confidence in how these systems make decisions.
This post builds on insights from our Agentic AI in HCM Trends Report to explore what that next stage looks like — and what leaders can do now to turn potential into performance.
Learn more: Opkey’s HCM Industry Trend Review & 2026 Forecast
From Assistants to AI Agents in HCM Workflows
In HR, as in other business functions, AI applications have generally developed in line with ISG’s three “eras” — from prediction to creation to orchestration. What began with systems that could forecast outcomes, such as attrition or hiring trends, soon evolved into tools that could generate content and automate simple workflows. These earlier waves of predictive and generative AI helped HR teams analyze workforce data, personalize employee communications, and automate routine tasks like scheduling or approval, but they largely operated as digital assistants rather than active participants in work.
With agentic AI, the era of orchestration is beginning to unfold. These systems can decide, adapt, and act within defined guardrails, coordinating across people, data, and applications to achieve desired outcomes. Workday describes this as AI that can “monitor systems, interpret real-time conditions, and initiate tasks across connected applications,” a model that moves HR closer to the idea of a truly intelligent, self-optimizing enterprise.
Look at any of the leading HCM platforms, and you’ll see this evolution playing out:
- Oracle Cloud HCM’s Journeys enable HR teams to guide users through complex processes like onboarding, benefits enrollment, or policy updates with step-by-step prompts that adapt to each role.
- SAP’s Joule agents assist with HR workflows such as compensation planning, talent development, and performance management — automatically surfacing candidate data, summarizing employee feedback, and initiating actions like creating job requisitions or scheduling check-ins.
- Workday’s Agent System of Record lets organizations deploy and manage fleets of AI agents that assist across HR functions. These role-based agents can automatically update employee records, verifying access permissions, and even suggest learning opportunities based on role or performance data, all within the same secure environment that governs human users.
This model is spreading far beyond these flagship platforms. In talent acquisition, analysts predict that AI agents will handle early-stage screening for roughly 30% of recruitment teams by 2028. And in HR and IT service delivery, agentic systems already resolve end-to-end employee requests without routing tickets to humans.
Each of these examples follows a common theme: AI moving from telling people what to do to doing it with them. In HR, that shift opens a new frontier, where intelligent systems no longer sit at the edges of cloud HCM but become active digital co-workers.
Agentic AI in HCM: What the Analysts Are Saying
As rapidly as the use cases for Agentic AI in HCM are multiplying, the path to maturity is uneven. Analysts agree that most enterprises are still in the early stages, experimenting with contained applications rather than deploying agents at scale.
Broadly speaking, Gartner predicts that agentic AI will autonomously resolve up to 80% of routine service tasks by 2029. That’s only a few years away, yet Forrester offers a note of caution:
“We are still in the early stages of agentic AI’s market impact; companies must test, learn, and iterate because these powerful systems can be misaligned, creating actions that are at best undesirable and at worst harmful to your customers and critical applications.”
McKinsey’s “One year of agentic AI” assessment echoes that sentiment, emphasizing that success depends less on the sophistication of the technology and more on how organizations adapt to it. Many companies still treat AI as an overlay on existing processes rather than rethinking how work itself is structured. As McKinsey notes, the biggest performance gains come when workflows are redesigned for collaboration between humans and agents — where each contributes to what they do best, while the system continually improves based on shared feedback loops.
This distinction matters for HR because it means progress will depend as much on governance and trust as on innovation. Most HCM systems today remain in an “operator” stage of maturity, capable of taking action but still reliant on human oversight. Advancing to the “actor” stage, and eventually to true autonomy, will require clear frameworks for transparency, accountability, and continuous learning.
Lessons From Deploying AI Agents in HCM
If maturity depends on trust and governance, early adopters are already showing what that looks like in practice. Across industries, pilot deployments of agentic AI reveal consistent patterns and practical lessons for HR and IT leaders.
1. Trust Comes First
Organizations that succeed with agentic AI start by making systems explainable and setting clear boundaries for autonomy. Workday’s own research shows that while nearly 75% of employees view AI agents as “important teammates,” only about 30% feel comfortable being managed by them. However, in deployments where agents’ roles and limits were clearly defined, trust rose from 36% among early experimenters to 95% among mature users.
Put another way, agentic AI earns adoption only when employees understand why an agent acted — and how that action fits into established business rules and human oversight.
2. Training Drives Impact
The biggest performance gains come when AI agents don’t just automate tasks but actively teach and support people. An EY survey found that while 84% of employees are eager to work with agentic AI, most feel undertrained. Yet adoption and productivity rise sharply in organizations that provide structured enablement and clear guidance.
When agents are designed to share knowledge and adapt alongside users — guiding them through new processes and reinforcing best practices — training becomes ongoing. That learning loop builds organizational resilience, ensuring every update or workflow change strengthens readiness rather than disrupting it.
3. Measurement Evolves
Scaling agentic AI demands a new way to measure success. As McKinsey notes in its one-year assessment, when companies deploy hundreds of agents, tracking only outcomes makes it nearly impossible to find the source of errors.
Instead, ongoing monitoring and evaluation must be built into workflows so teams can verify performance and refine logic in real time. IBM’s research reinforces the point: only 42% of process-oriented organizations have developed KPIs for agentic systems. Leaders measure entirely new forms of adaptability and impact, such as “reasoning coherence scores” or “decision accuracy rates,” to understand the true value of agentic deployments.
4. Workflow-First Design Delivers Results
In McKinsey’s observation, the greatest barrier to realizing agentic AI’s value is not technical, but structural. Many companies deploy agents into existing workflows and see only marginal gains. The real breakthroughs come when organizations redesign processes from the ground up to accommodate human-AI collaboration, with agents embedded as integral actors rather than add-ons. In these re-engineered workflows, agents expose bottlenecks, surface insights, and continuously improve how work gets done.
Notice the theme? Across these early deployments, agentic AI succeeds when it’s built for accountability, learning, and continuous improvement.
Continuous Readiness in the Workday World
Many HR leaders, uncertain about the reliability of agentic AI, will wonder what this looks like in practice. Opkey’s new Training Agent for Workday offers a glimpse into this new reality. Just released this month, the Training Agent demonstrates how agentic principles of autonomy, transparency, and continuous learning can move from theory into day-to-day operations.
Workday customers operate in one of the most dynamic enterprise environments. Quarterly updates, regulatory changes, and process refinements are constant. Traditionally, keeping up has required extensive manual retraining and documentation efforts that become outdated almost as soon as they’re deployed. The Training Agent changes that model, applying agentic principles to the everyday challenges of enablement and governance.
- Autonomous enablement: The agent automatically generates and refreshes training guides from validated test cases, ensuring every process update is accompanied by current, accurate instructions.
- Unified coverage: It spans both web and desktop workflows through Workday’s Companion App, so training content evolves alongside the entire system, not in silos. And because the Companion App runs directly within the Workday environment, users access guidance without leaving their normal workspace or disrupting daily tasks.
- Embedded compliance: Administrators can define access thresholds, and the agent continuously monitors for breaches during impact analysis.
- Personalized guidance: Contextual, role-based help appears directly in Workday to reduce support tickets and shorten the adoption curve.
This is what agentic AI looks like when it leaves the lab — a system that learns, adapts, and scales with the organization. More than an automation tool, it’s a blueprint for what maturity in AI-driven HCM really means: continuous readiness, trusted governance, and measurable impact at every release.
Building Trust for AI Agents in HCM
Agentic AI maturity ultimately depends on trust, and that starts with building stronger collaboration between humans and their growing base of digital co-workers. The organizations that lead this next wave of HCM AI agents will be those that embed accountability and transparency into every interaction, so people trust the agents working beside them.
Confidence grows when systems can clearly explain their reasoning, adapt to feedback, and facilitate better teamwork. In the coming year, platforms like Workday — and innovators like Opkey — will further enhance collaboration, embedding orchestration directly into agentic workflows to connect learning, compliance, and change management in real-time.
For a deeper look at how these dynamics are reshaping HCM, explore our full Agentic AI in HCM Trends Report.
