KOHb - Getty Images
Attention AI skeptics: The agent era has begun
Early AI limitations justified caution, but the rise of AI agents is changing the conversation. Act now to focus on practical adoption.
Are you an AI skeptic? It's a perfectly reasonable position, but -- as tempting as it is to rail against the hype -- it's time to change your mind. When ChatGPT arrived on the scene, it was impressive. Unfortunately, it also made mistakes, invented facts, and required patience and tenacity to be genuinely useful. It was fair enough to regard it as an interesting development, but not robust enough take seriously for real business use.
All the talk of an "AI bubble" raises questions as well. The five largest US cloud and AI infrastructure providers (Microsoft, Alphabet, Amazon, Meta and Oracle) have collectively committed to spending between $660 and $690 billion on capital expenditures in 2026, according to Futurum Research. That's nearly double what they spent in 2025. There's also a real gap between spending and returns. Hyperscalers have already spent nearly $400 billion this year in capital expenditure, while enterprise AI generated only $100 billion in actual revenue, private investment firm Cresset reported. These numbers make the bubble sound more like a bomb about to go off.
That huge flow of cash, however, is helping the technology advance at a breakneck pace. And unlike the dot-com era, which crashed due to debt-fueled speculation, today's AI investments are cash-funded by profitable businesses. Nvidia, for example, reported a record $57 billion in revenue for its fiscal Q3 2025, up 62% year over year. Demand for AI computing resources is far greater than supply, creating a much different situation.
Since the launch of ChatGPT in the fall of 2022, the usefulness and capabilities of AI have multiplied. What started out as an unreliable novelty, well-known for concocting advice like making pizza with glue, advanced faster than we could imagine. By late 2024, models like OpenAI's o1 introduced reasoning, meaning the AI could work through a problem and self-correct before giving an answer.
The recent emergence of widely available AI agents goes much, much further than answering questions. Agents can take on multi-step tasks, use tools, check their own work, and repeat the steps until the job is done. A human sets the goal; the agent executes it. This is functionally different from everything that came before.
A wave of AI-driven workforce restructuring is sure to follow. AI agents will let businesses reduce costs by eliminating some positions. But the biggest opportunity lies in their ability to dramatically increase the productivity of very small teams. One person directing multiple agents will be able to do work that previously required an entire department. Companies built with lean, "AI native" models will have both lower costs and more capability than traditional organizations with large overheads.
What IT leaders should do now
If you've been standing on the sidelines, waiting for AI to play out, now is the time to jump in. Agents may sound scary, but they also offer the ability to spin up new functions, departments, and products with far less risk. Companies that start experimenting and integrating these capabilities now will have a distinct advantage over those that hold back.
1. Pilot agentic tools, not just chatbots
Help your team get practical experience with tools such as Claude Code and OpenClaw. Ensure they start developing a deep understanding of the tools' potential and what's coming next.
2. Audit your workflows
Agents shine where work involves repetitive multi-step tasks, handoffs between people or processes that require checking and rechecking outputs. Map those workflows to steps an agent could perform.
4. Examine your staffing model
Think about how roles could shift and adapt. Positions that consist primarily of coordination, routing or repetitive tasks could be absorbed by agents. Have a game plan for redeploying people whose responsibilities are reduced to duties that can only be accomplished by humans.
5. Get ahead of governance now
Agents acting autonomously on systems and data will create new security, compliance and audit requirements. Define your acceptable use policies, access boundaries and human-in-the-loop requirements before deployment.
6. Push your executive team
The business case for AI investment is no longer speculative. Organizations that move slowly on agents will quickly be overtaken by competitors that are more agile. Ensure your company's board is aware of the stakes.
The progression of AI has moved from "cool but flawed" to "reliable" to "autonomous" in a surreal amount of time. If your AI strategy is still about chatbots and assistants helping individuals be slightly more productive, you're one stage behind.
The organizations and the IT leaders that will be successful aren't necessarily the ones spending the most money. They're the ones that figure out soonest how to direct agents effectively. It's a technology challenge, but even more so, it means rethinking how technology and humans work together and developing brand-new corporate ecosystems.
Now is the time to get serious about AI.
