AI to ROI: Big Story - The Great AI Jobs War

The Essential Resource for AI Leaders

The Great AI Jobs War: Hype, Hope, and a Few Hard Truths

Bottom line up front: AI isn’t coming for all jobs, but it is affecting which jobs exist and who gets them. CEOs have drunk the AI Kool-Aid and are promising their Boards and investors that major productivity gains are happening now or just around the corner. The reality is that AI-driven productivity gains are unevenly distributed and elusive.

Here’s one AI-driven jobs thing that’s certainly true:

Entry-level workers are getting crushed while electricians are making bank.

Beyond that, how AI affects the future of employment for American workers is anyone’s guess.

CEOs, the Boardroom, and Investors Have AI Fever, But Delivering on AI Promises Is Hard

CEOs can’t stop talking about AI:

• 306 of the S&P 500 companies mentioned AI on their Q3 earnings calls.

• By Q4, AI became the #1 topic for CEOs, appearing in 47% of all calls.

• Mentions of “agentic AI” and “digital labor” increased 779% year over year.

Here’s the problem: PwC’s 2025 Global CEO Survey found that:

• Only 10-12% of companies report meaningful AI benefits.

• A staggering 56% say they’re getting “nothing out of it.”

In addition, a deservedly maligned but still quoted MIT study pegged the failure rate of enterprise AI pilots at 95%.

Some mid-to-large-sized enterprises have been explicit in using AI to justify cuts to their workforces:

• Dow tied 4,500 layoffs to a plan “utilizing AI and automation” to improve shareholder returns.

• Pinterest cut 15% of its workforce while simultaneously admitting its platform had become “overridden with AI-generated slop.”

• Salesforce eliminated 4,000 customer support positions after claiming AI was completing 30-50% of the company’s workload.

• While 44% of employers report offering AI training programs, only 33% of employees confirm this.

• Gen Z workers’ confidence in their skills readiness plummeted 20 points to just 39%.

Forrester Research calls it what it is:

AI-washing, attributing financially motivated cuts to future AI implementation that doesn’t yet exist.

Back to all those AI mentions on earnings calls:

Most companies aren’t yet putting AI into production at scale, so all that AI trumpeting basically amounts to CEOs waving jazz hands to distract from their lack of progress.

There is no widespread evidence YET that AI consistently drives increased revenues and profits across industries.

Where the AI-Driven Productivity Gains Are Actually Showing Up

A few companies are proving AI’s value with cold, hard metrics.

Klarna is the flagship case study. Since Q1 2023, the fintech company has:

• Cut its workforce from 5,000 to 3,000 through attrition

• Doubled revenue and increased revenue per employee by 152% (from $575K to nearly $1M)

• Achieved 96% daily AI usage among employees

Insurance company Hiscox deployed Microsoft 365 Copilot to 3,000+ employees across 14 countries. Results:

• Claims processing dropped from 60 minutes to 10 minutes.

• Complex underwriting that took three days now takes three minutes.

The tools driving these gains fall into clear categories:

• Coding assistants (GitHub Copilot, Claude Code, Cursor) boost developer task completion by 55-81%. 41% of all code written in 2025 came from AI.

• Customer service AI is transforming contact centers. Bank of America’s Erica has handled 2 billion interactions, resolving 98% of queries within 44 seconds. Lyft cut resolution times by 87%. Gartner predicts AI will autonomously resolve 80% of common service issues by 2029.

• HR automation is accelerating rapidly. Industry analyst Bersin & Associates predicts a 30% reduction in HR staff by 2026, with AI “superagents” handling recruiting, onboarding, and employee services. Some 61% of HR leaders are now planning to deploy generative AI, up from 19% in June 2023.

• Enterprise AI assistants (ChatGPT Enterprise, Claude) can save workers 40-60 minutes daily. Heavy users report 10+ hours saved weekly.

OpenAI says 75% of enterprise users can now complete tasks they previously couldn’t perform. Anthropic’s Economic Index estimates widespread AI adoption could increase US labor productivity growth by 1.0-1.8 percentage points annually, doubling recent gains.

The Macro-Level View

Here’s where it gets complicated. The net job numbers don’t look catastrophic, but there are pockets of pain with more to come. First, the rosy macro-level predictions:

• The World Economic Forum projects that AI will create 170 million jobs while displacing 92 million by 2030, for a net gain of 78 million globally.

• ITIF data show that AI created roughly 10x as many jobs as it displaced in 2024.

• Yale Budget Lab found that occupational changes since ChatGPT’s launch are only 1 percentage point higher than during internet adoption, so “not out of the ordinary.”

These rosy polls hide some harsh realities.

Entry-level workers are getting hammered:

• Job postings for entry-level positions have dropped 35% since January 2023.

• Entry-level hires accounted for 25% of all IT hires in 2023; that share was 7% in 2025.

• Finance entry-level positions fell 24%.

• For the first time in 45 years, college graduate unemployment exceeds the national average.

• Anthropic CEO Dario Amodei warned that AI may eliminate up to 50% of entry-level white-collar jobs within five years.

• IMF Managing Director Kristalina Georgieva called it an “AI tsunami” coming for young people.

The structural problem:

AI excels at “drunt work” (digital grunt work), the exact tasks that junior employees are trained on, and becomes the foundation of experience that serves them well throughout their career. As a result, 35% of “entry-level” postings now require 2-3 years of experience.

Meanwhile, some sectors are booming because of AI. NVIDIA CEO Jensen Huang calls data center construction “the largest infrastructure buildout in human history.” McKinsey projects that, by 2030, the US will need:

130,000 more trained electricians

240,000 more construction workers

150,000 more construction supervisors

Salaries for AI-related construction positions have nearly doubled.

What the 2030 Workforce Could Look Like

The data points towards a pyramid inversion:

• Senior workers - needed for judgment, oversight, and AI orchestration - will become more valuable, and their ranks will grow (McKinsey projects +3.8 million higher-wage jobs) until they all retire - creating a knowledge crisis.

• Mid-level workers will be squeezed as organizations flatten and agents take on decision-making roles previously held by middle managers.

• Entry-level positions face the deepest cuts because AI performs well on entry-level tasks.

• Gartner predicts that by 2030, almost no IT work will be done by humans without AI assistance.

• 75% of white-collar jobs will be augmented by AI systems.

• The remaining 25%? Done by AI with little to no human intervention.

• McKinsey says workers in the lowest wage quintile are 10-14x more likely to need to change occupations than top earners.

“AI is coming for middle management first. Amazon’s latest restructuring aims to reduce organizational layers, and the number of managers at public companies dropped 6.1% from May 2022 to May 2025. The average supervisor now manages six direct reports, double the number five years ago. The question now isn’t whether AI will take jobs, it’s how you’ll protect yours.”

Scott Galloway on AI and middle management

The Honest Answer: Nobody Knows

Here’s what the conflicting data reveals:

AI’s employment impact depends on variables we can’t fully predict - adoption speed, regulatory responses, business model innovation, and human adaptability.

The optimists point to history. Technology has consistently created more jobs than it destroyed. 60% of US workers today are in occupations that didn’t exist in 1940. Marc Andreessen calls fears of AI unemployment a “fallacy.”

The pessimists counter: This time is different. We’re automating intelligence itself, not just physical labor. ChatGPT reached 100 million users in two months. The speed is unprecedented.

The realists land somewhere in between: The best guess is that two-thirds of jobs will experience partial automation - task-level change, not wholesale replacement. The winners and losers will be determined by industry, skill level, and adaptability. No matter what happens, the transition will be painful.

The Takeaways for CEOs and C-Suite Executives

• Audit the hype. If your AI investments aren’t showing measurable ROI, you’re in the majority. Concentrate on bringing your most promising projects to production and delivering real ROI.

• Maintain an entry-level pipeline. Today’s efficiency gains could become tomorrow’s talent crisis. Use AI to upskill entry-level workers faster.

• Place bets on augmentation. The biggest productivity wins come from humans + AI, not AI alone.

• Don’t ignore the trades. Physical work is suddenly a growth sector.

The AI jobs story isn’t about robots taking over. It’s about a massive redistribution of work - of skills, across the entire org chart, and for entire industries. Unfortunately, there’s no tried-and-true playbook yet, so you and your team will need to lean in and figure it out together.