What Is Agentic AI? White-Collar Robots Are Here, For Better or Worse

PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.

Big Tech has a new AI obsession that you can expect to hear a lot about in 2025: Agentic AI.

Though these "AI agents" are still in the early stages, the general idea is to move from a transactional relationship with an AI (user prompts, computer answers), to the AI being more autonomous. It can perform actions on your behalf, with minimal or no input from you.

A good introductory example is OpenAI's tasks, launched this week for ChatGPT Plus, Pro, and Team subscribers. The chatbot can now perform simple, routine actions, like checking the stock market or delivering a daily news summary every morning (better hope it doesn't hallucinate).

Some capabilities seem a little mundane, like setting a reminder for an appointment. (Don't we have calendars for that?). But the idea is to manage all of these tasks in one place. When it completes the task, a notification pops up on web, desktop, and mobile, The Verge reports.

This could be one reason OpenAI CPO Kevin Weill says that chatbots messaging humans first will be a "big theme" of 2025.

Businesses might be particularly interested in AI agents. They can boost productivity, or even replace some jobs entirely. Anthropic's "computer use," launched in October 2024, empowers the work-focused Claude chatbot to autonomously complete tasks on your computer, such as moving the cursor, opening web pages, typing text, downloading files, and completing other activities.

Around the same time, Microsoft launched a suite of 10 AI agents, each "fine-tuned," or customized, for specific job functions. "You can build a very rich agentic world defined by this tapestry of AI agents, that can act on our behalf across our working life, across teams, business processes as well as organizations," CEO Satya Nadella told Ignite attendees.

Today, Nadella announced , which adds pay-as-you-go agents to Microsoft 365 chat commercial customers. "At its core, AI empowers to ideate, to create, and to achieve our business goals," Nadella said in a video introducing Copilot Chat. "It gives each of us more agency and more leverage while freeing us from the drudgery of repetitive tasks."

Nvidia also debuted its AI agents for businesses early this year, calling them "knowledge robots" that can "reason, plan and take action." They are essentially white collar robots, replacing office work instead of the traditional way of imagining robots on, say, an assembly line.

This week, Nvidia expanded its agentic AI push into healthcare. It aims to "boost the efficiency of clinical trials and optimize planning for the launch of therapies and medical devices—ultimately improving patient outcomes." That's far more complex than early AI applications we've seen in healthcare, like transcribing patient notes and reading X-rays (sometimes with poor outcomes).

Are we comfortable with giving AIs this much control? We have reasons to be hopeful and skeptical. If the system works well, it could save time and be immensely helpful. But if it suffers from hallucinations, or performs an action incorrectly, the fallout could undermine any benefits and lead to more poor outcomes.

Like humans, AIs can also suffer from behavioral problems. In one study, researchers found that large language models (LLMs) built by OpenAI, Anthropic, and Meta were more prone to escalating violence in a war video game simulation than a human player. (These are real risks as both the China and the US race to master AI warfare.)

While you may be more likely to witness AI wreaking havoc on your computer than the battlefield, the more we rely on these systems the more important it becomes to validate they are performing as expected.