The new hybrid working model: Managing an agentic workforce

Agentic workforce: woman pointing and smiling at an open laptop in an office

AI has the potential to unlock more than $4 trillion in global productivity gains, according to McKinsey. It can make us more efficient, help us scale faster, and ensure our talent has ample career growth pathways.

Yet the ServiceNow Enterprise AI Maturity Index 2025 found that the average AI maturity score slipped. While executives are excited about the technology's potential to boost productivity across the enterprise, only about one-third of respondents have reached at least the piloting stage for any agentic AI use case.

The barrier isn’t technical. The models are good—and getting better. The barrier is structural. It’s mindset, leadership, and adoption.

It’s time to move the conversation from what could be to what is: The agentic workforce is here. It’s not in the future. It’s not in development. It’s in the enterprise today. Organizations pulling ahead in the AI race recognize this and are adapting the ways their businesses are structured, measured, and governed.

The agentic workforce, deployed

The agentic workforce is comprised of specialized AI agents that are embedded in roles across functions. These AI agents automate tasks, yes. They also take on defined roles, complete workflows, make decisions, and improve with feedback.

AI agents are part of the broad team that makes up the enterprise. They’re a core part of how business results are delivered.

That said, like any part of a team, they require structure, oversight, and direction. Although they’re not “co-workers” you’ll find at the watercooler, they still need managers.

The organizations pulling ahead aren’t waiting for a perfect use case. They’re building the infrastructure, accountability, and leadership mindset to support agentic teams at scale.

The shift from tools to team

In this new hybrid working model, there’s no such thing as an individual contributor anymore. Every contributor is going to be managing a team of agents.

At ServiceNow, we’re handing off decisions to our AI models when reliability, risk, and oversight permit. This isn’t about an intriguing conversational AI agent with a great user interface. We power complex businesses. Data integration that takes into account multiple systems and workflows is a must.

Taking this approach, we’ve reached a fully agentic autonomous IT service desk. We’ve long said that when 80% of a role could be done by the agentic workforce, talent and resources would be freed to work on more strategic efforts. We’ve blown past that goal and have reached 90% of IT support work fulfilled by AI agents.

We have not, as popular media would suggest, sent our service talent to the unemployment line. We’ve redeployed those workers to help our company grow. The company is better off, and the talent is happier.

In addition, since deploying our agentic workforce:

Legal support requests have been resolved 80% faster. This has saved hours for our highly valuable legal talent, freeing teams to focus on more complex issues.
Security risk assessments are 66% more efficient. More than half of false positive phishing cases are resolved within 20 seconds, and incidents are closing seven times faster than without AI agents, significantly cutting downtime and business disruption.
We’ve reached 99% speed gains on our sellers’ commissions questions, nearly eliminating the need for request tickets. Instead of waiting up to four days, sellers can get answers in eight seconds, freeing them to focus on talking to customers.

This transformation is happening across sectors.

In healthcare, AI agents support scheduling and billing.
In finance, they streamline client onboarding and monitor risk.
In logistics, they rebalance inventory and reroute shipments in real time.

Governance and risk: Constraint to accelerator

The agentic workforce is powerful—but without structure, it’s also risky. Governance can‘t be an afterthought. It must be the foundation. If we expect these systems to carry business-critical responsibilities, we must treat them with the same care and discipline we apply to human teams.

Poorly governed AI agents won’t fail dramatically; they’ll degrade quietly. They won’t call in sick, but as data shifts and conditions evolve, performance can slip. That’s why regular feedback loops, monitoring, and retraining are essential.

At ServiceNow, we’ve built AI Control Tower to help manage exactly that. It gives managers full visibility into which agents are active, what they’re doing, and how they’re performing. Metrics such as throughput, accuracy, sentiment, and system health are tracked.

Managers can see when retraining is needed, when escalation is appropriate, and when agents are no longer adding value. This isn’t a tech feature. It’s workforce management. As on any team, quality doesn’t come from “set it and forget it.” It comes from leadership.

Today, oversight might live with chief AI or analytics officers. In time, however, AI agent management will be part of every manager’s daily workflow, just like checking team dashboards or reviewing key performance indicators. Agentic performance will become part of our operating rhythm.

At ServiceNow we’ve long said that when 80% of a role could be done by the agentic workforce, talent and resources would be freed to work on more strategic efforts. We’ve blown past that goal and have reached 90% of IT support work fulfilled by AI agents.

Slow adoption isn’t conservatism. It’s risk.

The urgency isn’t just technical—it’s strategic. Deloitte found that an incredible “89% of CEOs are exploring, piloting, or implementing agentic AI within their organizations.” Yet many organizations haven’t built the foundational governance or enterprise strategy to scale AI safely. They haven’t started to address the mindset needed in their company cultures, either.

MIT’s State of AI in Business study got a few things right: The majority of businesses often try implementing AI without cohesion, get stuck in experimentation or limited deployment, and have poor integration of AI tools into their company workflows.

Each of those scenarios is avoidable when leadership steps up. Many businesses are limiting themselves to easy use cases to drive quick value—but not transformation.

There are no shortcuts

This is not a case of “think bigger.” It’s a case of “do bigger.” It takes time and resources to deconstruct what each department and role does on a day-to-day basis and then reimagine it with the agentic workforce in mind. It’s time for leaders to see the possible, build alignment, and make it happen.

That said, business transformation isn’t only top-down work. The bottom-up work is building and embracing an AI culture for career growth. Talent will always play a critical role—defining responsibilities, training the systems, reviewing outputs, and stepping in when performance changes. Talent that isn’t open to AI or skilled in the tools may get left behind.

This isn’t a call for rash and hurried adoption. It’s a call for deliberate readiness. The organizations pulling ahead aren’t waiting for a perfect use case. They’re building the infrastructure, accountability, and leadership mindset to support agentic teams at scale.

They’re picking a strategic metric that will boost their top line and reduce friction internally—such as revenue per employee—and running at it. They’re treating AI as a fully embedded part of their operating model, not simply a tool. They’re remembering the goals:

Stop using employees’ time for lower-order, repeatable tasks.
Build capacity and operational efficiency.
Put AI to work for people.

The organizations defining the next era are building and embracing agentic AI that delivers. That’s how we power the future of work.

Find out how ServiceNow can help you put AI agents to work for people.