This week brings clarity on how AI is restructuring work — who is being cut, who is still advancing, and where demand is shifting. New data shows flattening hierarchies, automation at scale, and early evidence that AI adoption is seniority biased. At the same time, freelance markets and consumer surplus numbers point to hidden gains, and a large hiring study suggests that your best chance of landing a job may come from talking to an AI.
What to Know:
Now that AI is quietly taking over routine approvals, scheduling, and performance tracking, companies are eliminating layers of middle management to cut costs and flatten hierarchies while justifying larger reporting spans.
The median manager-to-employee ratio has tripled since 2017, rising from 1:5 to about 1:15, with firms such as Google cutting 35% of small-team managers, Intel eliminating half of its management layers, and Citi, Amazon, and Bank of America also reducing supervisors. Managers now juggle bigger teams with less time for coaching or mentoring, leaving employees to self-promote or seek guidance from peers.
Why It Matters:
Flattening is reshaping workplace culture. With fewer managers, only highly independent workers thrive, while career development and personal support decline. Last year, I interviewed Gartner's Vice President Analyst, Chief of Research Daryl Plummer on his prediction that by the end of 2026 20% of enterprises will have flattened. This is likely an early sign of that.
As hierarchies collapse, AI isn't just changing how many managers there are — it's also starting to replace entire job categories.
What to Know:
Marc Benioff confirmed Salesforce cut about 4,000 customer service roles, reducing headcount from 9,000 to 5,000, as AI agents now handle half of all customer interactions. The shift has lowered support costs by 17% since early 2025. Klarna and Microsoft have made similar cuts, and Benioff said he is reviewing "every single function" at Salesforce for potential automation.
Why It Matters:
Despite earlier reassurances, Salesforce now openly acknowledges replacing workers with AI. The move signals accelerating automation in sales and support sectors.
To understand who is actually being displaced, two major studies provide the clearest evidence yet.
What to Know:
Using data on 62 million workers and 285,000 firms from 2015 to 2025, Harvard researchers found that after early 2023, firms adopting generative AI cut junior employment by 7% to 12% while senior employment kept rising. The decline was driven by slower hiring, especially in wholesale and retail, not layoffs. Promotions of juniors already inside firms increased, but new entry slowed sharply. Effects varied by education: Mid-tier university graduates faced the steepest declines, while elite and lowest-tier graduates were less affected.
What to Know:
Using payroll data from ADP covering millions of U.S. workers, researchers found that employment for young workers (22 to 25) in AI-exposed jobs, such as software development and customer service, fell 13% since late 2022, while older workers in the same jobs and workers in less-exposed fields saw stable or rising employment.
A Stanford University study showed that in jobs where AI is used to augment work, entry-level employment does not decline — in fact, those occupations saw stable or even growing hiring compared to automation-heavy ones
Why They Matter:
Both studies found that entry-level workers are bearing the brunt of AI-driven disruption. Automation is erasing traditional on-ramps into exposed professions, even as augmentation leaves opportunities intact for more experienced staff. This "seniority bias" risks narrowing pathways into firms, with long-term consequences for wage growth, mobility, and inequality.
The Stanford study used payroll data to show early-career job losses in exposed occupations.
The Harvard study used resume and posting data to show within-firm seniority bias.
It's easy to focus only on cuts and displacement, but parallel signals show demand reshaping in real time.
What to Know:
In August 2025, high-value freelance contracts ($1,000+) from large firms rose 31%, with one in four leaders hiring freelancers for advanced technical skills. Demand for AI and machine learning talent jumped 40% among small and medium-sized businesses, while creative categories such as video, content writing, and language tutoring also grew sharply. Fact-checking and video editing ranked among the top AI-related skills, underscoring the need for human oversight of AI outputs.
Why It Matters:
Freelance markets are an early signal of how AI is reshaping demand. Companies are using AI to speed work but leaning on freelancers for judgment, creativity, and technical integration.
And while firms are still recalibrating their labor mix, and AI productivity isn't showing in widespread macro-economic data, the benefits for individuals are already visible at scale thanks to a clever study from Erik Brynjolfsson at Stanford.
What to Know:
Erik Brynjolfsson, a professor at Stanford University and senior fellow at the Stanford Institute for Human-Centered AI, and Avinash Collis, an assistant professor at the Heinz College of Information Systems and Public Policy at Carnegie Mellon University, estimated that Americans gained $97 billion in consumer surplus from generative AI in 2024 — benefits such as saved time and free creative tools were not captured in GDP. Their survey found regular users would need $98 per month in compensation to give up AI, with 40% of U.S. adults using it regularly. The authors argued GDP lags because adoption costs show up before complementary investments pay off, producing a "productivity J-curve."
Why It Matters:
Official GDP data understates AI's impact. While firms are still adapting, consumers have already captured large welfare gains — especially lower-income households who benefit most from free tools.
Those same benefits extend beyond convenience — early trials suggest AI may even improve the odds in hiring.
What to Know:
A field study of 70,000 applicants in the Philippines found that candidates interviewed by AI agents were 12% more likely to receive job offers and 17% more likely to stay at least one month. Nearly 78% of applicants opted for AI-led interviews when given the choice, with 70% rating the experience positive. However, recruiters needed about twice as long to review AI-generated notes compared to human-led sessions.
Why It Matters:
The research suggests AI can improve hiring outcomes and applicant satisfaction, but efficiency gains for recruiters remain uncertain.