AI Agents Are Moving From Experiments to Operating Models

In recent years, the discourse surrounding artificial intelligence (AI) in executive circles has predominantly revolved around tools—specifically, which AI copilots to adopt, which large language models to implement, and how to manage the risks of shadow AI. However, as organizations begin to mature in their AI journeys, the conversation is shifting towards a more profound transformation: the emergence of AI agents. These systems are capable of completing tasks, coordinating workflows, and interacting with other systems, all while requiring less direct human intervention. This transition signifies not just a technical evolution, but a fundamental shift in operating models within organizations.

The Transition to AI Agents

AI agents differ from traditional AI assistants in their operational capabilities. While AI assistants typically respond to user prompts to perform tasks such as drafting emails or summarizing meetings, AI agents can autonomously monitor inputs, trigger actions, and manage workflows without waiting for human input at every step. This shift is significant because it allows organizations to rethink how work is conducted, moving beyond mere automation to a more integrated approach where humans and machines collaborate effectively.

According to Deloitte, by the end of 2025, 25% of companies utilizing generative AI are expected to initiate pilot programs for agentic AI, with this figure rising to 50% by 2027. McKinsey's global AI survey indicates that 23% of organizations are already scaling agentic AI systems, while another 39% are still in the experimental phase. Despite this momentum, the majority of companies remain in the early stages of adopting AI agents, highlighting a gap between experimentation and practical implementation.

This gap is critical as it represents a pivotal moment for organizations to either evolve their operating models or risk stagnation. Many companies currently view AI agents as mere productivity enhancements—such as a customer support agent or a sales research assistant—rather than as integral components of their decision-making and operational frameworks. This fragmented approach can lead to confusion and inefficiency, as AI agents may exacerbate existing issues within the organization rather than resolve them.

Redesigning Workflows with AI

The key to successfully integrating AI agents lies in understanding the workflows that require human judgment and where decision-making processes may falter. Leaders often mistakenly ask, "Where can we use agents?" instead of focusing on where work consistently stalls due to breakdowns in judgment, context, or coordination. This shift in questioning is essential, as it guides organizations to identify critical areas where AI can truly add value.

To harness the full potential of AI agents, leaders must cultivate a robust Judgment System—an organized method for capturing, synthesizing, deciding, and acting with AI as a collaborative partner. At the organizational level, this translates into what is known as Judgment Infrastructure, which encompasses the people, processes, and AI systems that enhance decision-making quality as speed increases across teams. This infrastructure becomes increasingly vital as AI agents expedite actions; without proper judgment, rapid decisions can lead to chaos, while slow decision-making can render organizations ineffective.

The companies that will thrive in this new landscape are not necessarily those that deploy the most AI agents, but those that effectively redesign their workflows to improve decision velocity without sacrificing the quality of those decisions. AI agents should not be dispersed throughout the organization as isolated experiments; instead, they should be strategically integrated into recurring decision-making processes to enhance overall efficiency.

The Importance of Cultural Foundations

The successful implementation of AI agents also hinges on the cultural environment within an organization. Leaders must foster a culture of psychological safety, where employees feel secure in sharing their experiences with AI and are encouraged to experiment with new workflows. This cultural foundation is crucial, as it prevents employees from viewing AI as a threat to their jobs and instead positions it as a tool to amplify their skills.

Case studies from organizations like Progyny illustrate the transformative potential of AI agents when aligned with a supportive culture. For example, Progyny's CEO, Pete Anevski, utilized AI to enhance the clarity of 1:1 meetings, leading to shorter decision cycles and improved accountability. Similarly, the HR team at Progyny developed an AI bot to streamline responses to common employee inquiries, which not only expedited answers but also provided valuable feedback on areas needing improvement.

In contrast, leaders who project an image of certainty without embracing the learning journey can hinder the potential of AI agents. As seen in the experience of Skyscanner's CTO, Andrew Phillips, openly sharing challenges and lessons learned fosters a collaborative environment where teams can adapt and innovate together.

As organizations navigate the next phase of AI adoption, the focus should not solely be on the sophistication of their agent strategies but rather on the clarity of their judgment infrastructure. The ability to effectively integrate AI agents into workflows will ultimately determine an organization's readiness to leverage these powerful tools. The question remains: Is your organization prepared to evolve alongside AI agents?