Last month, McKinsey released its most recent report on AI agents titled What Is an AI Agent? Given the focus on tariffs dominating the news cycle, it’s not surprising if you missed the report, but it comes at an important time in the growth of AI. AI remains a structural growth story regardless of tariffs, and AI agents are accelerating in usage by the day. According to the Cloudera Enterprise AI Maturity Report 2025, 96% of enterprises across the globe plan to expand their use of AI agents in the next 12 months, with more than half aiming for organization-wide adoption. The Mckinsey report offers a compelling blueprint for understanding how AI agents are transforming enterprise productivity, digital workflows, and the nature of decision-making itself. These agents—software components capable of autonomously acting on behalf of users—represent a shift from passive assistance to proactive orchestration. As Bill Gates noted, “Agents are smarter. They’re proactive—capable of making suggestions before you ask for them. They accomplish tasks across applications. They improve over time because they remember your activities and recognize intent and patterns in your behavior.” Gates describes the proactive nature of agentic AI, illustrating how these systems can anticipate user needs and evolve through interaction, potentially revolutionizing user experiences.
In other words, given enterprises have been slow to adopt copilot-style AI tools, AI agents will speed up adoption—leading to faster efficiency and productivity. They can now plan, execute, and improve upon complex tasks, often coordinating with other agents or humans in the process. The McKinsey paper serves as the intellectual springboard for this broader reflection on where AI agents are headed, how quickly they’re being adopted, and why their most immediate and profound impact may be felt in healthcare.
While AI agents have existed in early forms for years, we are now witnessing the industrialization of agentic systems. The past several months have marked a turning point in real-world deployment. In April, OpenAI launched Operator, a fully autonomous agent designed to perform multi-step tasks without constant human supervision. Operator doesn’t just respond to prompts—it books meetings, files reports, completes purchases, and can run entire workflows. Alongside Operator, OpenAI introduced a robust toolkit: the Agents SDK and Responses API, allowing developers to build customized agents capable of interfacing with apps, APIs, and databases. Google, meanwhile, announced its own comprehensive agent strategy at Cloud Next 2025. Centered on the Gemini 2.5 model, Google’s stack includes Workspace integrations, the AgentSpace platform, and support for agent interoperability protocols like Anthropic’s Modal Context Protocol (MCP). These innovations position Google not just as a model provider but as a full-stack enabler of autonomous systems. Anthropic, too, continues to push boundaries. With offerings like Claude Code and MCP, it is making software generation and enterprise automation accessible to non-technical users, reinforcing the accelerating shift from AI that simply answers to AI that autonomously acts.
This shift is not a niche movement. Industry adoption is exploding. By the end of 2025, it’s projected that already 85% of enterprises will be using AI agents. The global AI agent market, which stood at $5.4 billion in 2022, is expected to grow to $47.1 billion by 2030. That exponential growth is powered by foundational advances in natural language processing, reasoning, multimodal learning, and the integration of agents into enterprise architectures. As Eric Schmidt recently observed during a 2024 policy hearing, “Inference and reasoning are where the real breakthroughs will come from. That’s when AI agents begin to understand goals, consequences, and multi-step decisions—not just generate text.”
These capabilities are already being woven into the fabric of consumer devices. Apple has received criticism for its “Apple Intelligence” initiative so far, but together with Samsung’s Galaxy AI platform, they are introducing AI-native smartphones—devices where AI is no longer a separate app but the central nervous system of the phone. Each release will improve, and by early 2026, it is expected that consumers will begin to see the agentic benefits. These phones will process data on-device, invoke AI agents seamlessly across apps, and enable real-time reasoning with full user context. Eric Schmidt called this the next frontier: “AI agents will be the new apps, and your phone will be their command center.” Just as the iPhone redefined the smartphone in 2007 by embedding the internet into your pocket, AI-native devices will redefine personal computing by embedding decision-making power directly into the OS.
But nowhere are AI agents more urgently needed—or more promising—than in the healthcare sector. If you sit back and think about your experiences with healthcare—whether for yourself or your family—it’s striking how inefficient and time-consuming it has become from beginning to end. Healthcare systems across the world face converging crises: rising patient demand, long wait times, staffing shortages, and physician burnout. The situation is only worsening with demographic pressures and global health challenges. In this context, AI is not a “nice to have.” It is the only viable way to scale human care while reducing cost, complexity, and administrative overload. According to the 2025 Global Enterprise AI Survey, 94% of healthcare organizations now view AI as core to operations, and 86% are already using AI extensively. The global healthcare AI market is projected to surpass $120 billion by 2028.
Healthcare leaders increasingly believe the future lies in agentic AI—systems where autonomous agents complete documentation, schedule appointments, update records, and manage communications, all while ensuring that clinical staff can focus more of their time on direct patient care. AI agents are now handling routine yet critical tasks like data entry, prescription tracking, intake forms, billing cycles, and waitlist management. These systems are not just reducing errors—they are cutting costs and reshaping the patient experience. In the U.K., Guy’s and St. Thomas’ NHS Trust deployed enterprise agents to detect and correct waiting list errors, saving time and freeing up human staff to focus on care. Banner Health, one of the largest nonprofit hospital systems in the U.S., used digital agents to migrate millions of electronic medical records—an initiative that returned 1.2 million hours to the business. To see the benefits as an investor just look at the performance of the names within the S&P 1500 Healthcare Provider index.
AI is also being used to transform pharmacy logistics, regulatory compliance, telemedicine, cancer screening workflows, and even supply chain management. AI-enabled chatbots and virtual assistants are helping hospitals improve communication with patients, automate reminders, and streamline appointment scheduling. Portsmouth Hospitals, for example, increased maternity appointment capacity by 33% while saving over £100,000 in outsourcing costs—all with AI automation.
The state of AI agents is evolving rapidly, and the need for strategic alignment is urgent. McKinsey’s report notes that the most transformational value won’t come from standalone tools but from orchestrated systems where agents work together. This emergent behavior—where capabilities develop not from explicit programming but from dynamic collaboration—is the true promise of agentic AI. And in a field like healthcare, where the stakes are high and resources are strained, these agents will not simply change workflows—they may redefine what is humanly possible. If you are not convinced, just look at a company like McKesson and see their annual 20+% returns for the last seven years, including during the tariff panic of 2025!
