64% of Companies Just Changed How They Hire Entry-Level Workers. AI Is Why.

KPMG's Q4 AI Pulse Survey dropped a number that should make every hiring manager pause: 64% of organizations have altered how they hire entry-level workers because of AI agents. That's up from 18% just one quarter ago.

Three months. A 46-point swing in how companies think about their youngest hires.

This isn't gradual change. This is a structural shift happening in real-time.

What's Actually Changing

The traditional entry-level playbook looked like this: hire smart generalists, train them on your systems, and let them grow into the role over 12-18 months. The first year was essentially an investment—you paid for potential, not output.

AI agents broke that math.

When an AI employee can handle the tasks you used to give junior hires—data entry, research compilation, scheduling coordination, basic customer inquiries—the value proposition of "smart person who needs training" changes dramatically.

Companies aren't necessarily hiring fewer people. They're hiring different people.

The New Skills That Actually Matter

Here's what the survey reveals about which candidates are getting callbacks in 2026:

AI collaboration fluency. Not "knows how to use ChatGPT." Real fluency means understanding how to break complex tasks into agent-friendly chunks, when to trust AI output versus verify it, and how to debug when an agent goes sideways. Companies are asking candidates to demonstrate this in interviews—giving them a task and an AI tool and watching how they work together.

Process design thinking. Entry-level roles used to be about executing processes. Now they're about improving them. When the repetitive work is automated, what's left is identifying inefficiencies, spotting edge cases the AI misses, and redesigning workflows. Candidates who can think in systems have an edge.

Human-to-human skills. Sounds counterintuitive, but as AI handles more routine communication, the strategic human interactions become more valuable. Negotiation, relationship building, handling emotionally complex situations—these are the tasks AI employees still can't match.

Domain expertise over generic skills. "I'm a fast learner" used to be a selling point. Now companies want people who already understand their industry. Why? Because onboarding an AI employee is easy. Onboarding a human who can supervise that AI and catch industry-specific mistakes takes someone who knows the territory.

The "AI-Native" Advantage

There's an emerging divide between candidates who grew up treating AI as a tool and those scrambling to learn it now.

AI-native workers—mostly 2024-2026 graduates—have a different relationship with these systems. They've written papers with AI assistance, debugged code with AI pair programmers, and managed projects using AI tools. They don't think of AI as a separate capability to learn. It's just how work gets done.

For employers, this translates to faster ramp-up times. An AI-native hire doesn't need training on "how to use AI in your workflow." They need training on your specific business context. Big difference.

One CTO I spoke with put it bluntly: "We used to hire for raw intelligence and train the tools. Now we hire for tool fluency and train the context."

How SMBs Should Rethink Their Hiring Funnel

Large enterprises can afford to maintain traditional hiring while experimenting with AI. Small and mid-sized businesses don't have that luxury. Here's how the smart ones are adapting:

Redefine the job before you post it. Before writing a job description, audit what an AI employee could handle. What's left? That's the actual role. Too many SMBs are still posting descriptions for jobs that are 60% automatable, then wondering why their new hires seem underutilized.

Test for collaboration, not just competence. Add an AI collaboration component to your interview process. Give candidates a real task, give them access to AI tools, and evaluate the result. You'll learn more in 30 minutes than you would from an hour of behavioral questions.

Hire for the work AI can't do. If you're bringing on an entry-level customer success rep, don't evaluate them on ticket resolution speed—an AI handles that. Evaluate them on de-escalation skills, empathy, and creative problem-solving. Those are the moments that build customer loyalty.

Consider hybrid roles from day one. The old model: hire a person, then maybe give them AI tools. The new model: design a role that's explicitly human + AI from the start. The job isn't "marketing coordinator." It's "marketing coordinator managing three AI agents for content, scheduling, and analytics."

The Uncomfortable Truth

Here's what nobody wants to say out loud: some entry-level roles shouldn't exist anymore.

Not because the work isn't valuable. But because the work has changed. A company that hired five junior analysts to wrangle spreadsheets in 2023 might need two senior analysts who can manage AI-powered analysis in 2026. The output is better. The headcount is lower. The skill requirements are higher.

This isn't a reason to panic. It's a reason to adapt.

For job seekers: the bar is higher, but so is the leverage. If you can demonstrate real AI fluency—not just "familiar with AI tools" but actual collaborative skill—you're competing in a smaller pool.

For employers: the candidates you want exist. But you need to know what you're looking for. "Entry-level" doesn't mean what it meant two years ago. Update your expectations, your job descriptions, and your interview processes accordingly.

The 64% of companies already making these changes aren't ahead of the curve. They're responding to a reality that's been building for 18 months.

The question is whether you're in that 64%—or still hiring like it's 2023.