While many organisations are still grappling with the widespread adoption of generative AI, the next leap forward in the technology is coming fast – and it promises to usher in even more disruption for managers and HR departments.
Agentic AI broadly describes enabling AI-powered systems to speak to each other in a common language, so they can execute entire end-to-end processes that require collaboration and the application of logic.
While the parameters and outcomes might be defined by humans, AI ‘agents’ will make decisions and carry out transactions on their own initiative, leading Microsoft to claim employees will effectively be ‘bosses’ commanding a small army of bots. Others predict all-human workforces will soon be consigned to the history books, as smart AI does the grunt work in everything from financial services to retail.
How companies are already using agentic AI
Examples of agentic AI already taking place include mortgage lenders using agents to summarise applicants’ credentials and make recommendations on whether to lend to them; agents monitoring multiple CCTV feeds and deciding when to escalate incidents; and AI ‘personal shoppers’ who find the right online products based on customers’ preferences and budgets.
Documented users of agentic AI include everyone from Renault Group and Freshfields, a law firm, to Bayer and Gymshark. What such agents all have in common is the ability to use tightly honed intuition to improve their decision-making. By bringing together information from different sources and making value judgments on it, they are more like a human than an existing model, such as ChatGPT, which answers more linear questions.
Historically, people have been rewarded for management complexity but we will see a move towards skills-based organisations
“If there’s a piece of work human beings can do, even when it is working across myriad different systems or even physical pieces of paper, agentic gives software agency to do that,” says Prasun Shah, global CTO and AI lead in workforce consulting for PwC.
The logical next step, says Shah, is for such agents to interact more deeply with humans. He predicts the rise of bespoke personal assistants who complete everyday tasks on behalf of employees with minimal input. Shah adds that some businesses are also experimenting with digital twins, which shadow staff in their everyday tasks to learn how to operate on their behalf.
Microsoft, perhaps predictably, is even more evangelical about the technology. Spokespeople for the company have stated that they believe humans will “amplify their impact” and “think like a CEO” by becoming agent-bosses directing AI systems to do their bidding. As one spokesperson put it: “Agent-bosses don’t just do work – they orchestrate it”.
The dangers of moving too fast on agentic AI
Most businesses today are a long way from that point. According to a recent report from Mclean & Company, an HR research consultancy, only 7% of HR leaders globally say their business has a documented AI strategy and many are struggling to work out when and if staff should use the technology. In many cases, piecemeal adoption of AI technology is causing internal schisms.
Despite these issues, agentic AI is on the march. By 2030, says Accenture, more AI agents will be using ERP software than humans. “We are definitely expecting 2026 to be the year of agentic,” says Shah. “But it won’t stop there because models will evolve and you will see a move away from big, fat ERPs and see agentic-led business models emerge. More and more sophisticated AI employees will be created in a way whereby you can assemble and disassemble them to create more complicated AI personas.”
Shah says one of his clients is developing an agentic negotiating panel for its sales teams, where AI plays the role of the procurement professionals they’ll be pitching to, enabling them to hone their strategies with counterparts who react in real time to new information.
There is a danger of moving too fast in this area, however. In early 2024, payment provider Klarna said its new AI-powered assistant was doing the work of 700 customer service staff. In May 2025, it said it was hiring again because customers found the AI interactions unsatisfactory. Deployed correctly, however, the appeal of turbocharging human performance is irresistible. By 2030, a Salesforce study says 80% of business leaders believe they will have an AI agent in their ranks. Salesforce’s UK CEO Zahra Bahrololoumi has said: “I believe business leaders today are the last generation that will lead an all-human workforce.”
The impact of agentic AI on HR
This has multiple implications for HR. Kirk Chang, professor of human resource management and technovation at the University of East London, says human employees will be required to define boundaries, set goals and monitor AI outputs. They will focus on mentoring and innovation, while AI does the analysis in the background.
“To harness agentic AI’s full potential, HR must ask critical questions about ethics, readiness and governance and invest in upskilling and structural adaptation to ensure AI augments rather than undermines core HR functions,” he adds. “With thoughtful implementation and a focus on ethical governance, agentic AI can help HR leaders reimagine their function and drive organisational success.”
With thoughtful implementation, agentic AI can help HR leaders reimagine their function
Chang says HR is a powerful test case for agentic AI since so many of its core functions, including performance management and recruitment, are underpinned by data-driven processes. But HR must also be involved in any agentic AI rollout across the organisation: he advises HR leaders to audit the current readiness of their business to adopt new technology both technically and culturally, to get advice on the new skills and competencies required within HR teams to manage, interpret and oversee AI-driven processes. They should also consider the governance structures needed to monitor AI decisions and intervene when necessary.
Senior leaders, including in the HR function, should practice using AI tools, he adds, since many are less well-versed in the practicalities than their team members. Shah, meanwhile, says it’s time to consider the type of structure and future staffing you’ll need when agentic AI joins the workforce: “When work is taken over by agentic AI, inevitably you will see organisational structures being flattened and the focus will shift to deep specialisms and skills inside an enterprise. Historically, people have been rewarded for management complexity but we will see a move towards skills-based organisations.”
There will be winners and losers in the agentic AI revolution and there are plenty of different predictions for what happens to human workforces if adoption becomes widespread. Whether we will all be commanders of our own agentic armies very much remains to be seen.
While many organisations are still grappling with the widespread adoption of generative AI, the next leap forward in the technology is coming fast – and it promises to usher in even more disruption for managers and HR departments.
Agentic AI broadly describes enabling AI-powered systems to speak to each other in a common language, so they can execute entire end-to-end processes that require collaboration and the application of logic.
While the parameters and outcomes might be defined by humans, AI ‘agents’ will make decisions and carry out transactions on their own initiative, leading Microsoft to claim employees will effectively be ‘bosses’ commanding a small army of bots. Others predict all-human workforces will soon be consigned to the history books, as smart AI does the grunt work in everything from financial services to retail.
