September hiring all but stopped, and AI's benefits went to workers who already had good jobs
September payrolls came in at 29,000 with July revised to a loss, three companies filed AI-linked job cuts in the same week, and two new studies show AI use and its reported benefits landing on the workers who already had the better jobs.
Coverage window: October 2 – October 8, 2026
American employers added 29,000 jobs in September, the Bureau of Labor Statistics reported on 2 October, and the agency revised July's gain of 21,000 down to a loss of 10,000. Over the previous twelve months, payrolls had grown by an average of 45,000 a month. 1
Three companies filed workforce reductions in the same week. HubSpot told the Securities and Exchange Commission on 6 October that it would cut about 660 roles, 7% of its staff, as it shifts toward AI-driven customer outcomes. FICO filed the same day to cut about 15% of its staff and fold AI-driven product development into a flatter organization. Amazon said on 6 and 7 October that it had eliminated fewer than 1,000 roles, most of them in the retail business, alongside a planned $220 billion of spending on data centres, chips and other AI infrastructure. 234
Two studies published in the same days describe who inside a company actually reaches AI, and what happens to hiring when they do. Gallup found that 28% of American employees use AI at work at least weekly and 54% never use it, with regular use running at more than twice the rate among college graduates as among workers without a degree. The Federal Reserve Bank of New York found that small firms using AI expect to add staff at more than twice the net rate of firms that do not. 56
Set the 29,000 against those two findings, and the week's question for anyone running a team becomes where the work actually goes.
| Signal | What changed this week | What to watch |
|---|---|---|
| September hiring | Employers added 29,000 jobs, and the agency revised July down to a loss of 10,000; unemployment held at 4.2%. 1 | Whether the next report runs below the 45,000-a-month average of the past year, or the revisions were noise. |
| AI-cited cuts | HubSpot cut about 660 roles, FICO about 15% of its staff and Amazon fewer than 1,000, all filed between 1 and 7 October. 234 | Whether each company's own explanation of the change matches the framing of its headline. |
| Who gets AI | 28% of employees use AI at work at least weekly and 54% never do; college graduates use it at 40% against 17%. 5 | Whether access widens past managers and graduates, or stays where it began. |
| AI-adopting small firms | 46% of employer firms use AI; adopters expect headcount to rise by a net 33 points against 15 for firms that do not. 6 | Whether those expectations turn into hires, which the survey does not yet measure. |
| Where the work moved | Amazon was recruiting former employees for AWS and AI roles in the same week it cut retail jobs. 4 | Whether other large employers publish a route from the roles they close to the roles they fill. |
September's hiring all but stopped
Payroll growth of 29,000 in September compares with the 45,000-a-month average of the prior year, and the annual revisions took 60,000 jobs off July and August together: July now shows a loss of 10,000 and August 133,000. 1
The unemployment rate held at 4.2%, inside a band of 4.1% to 4.3% that has held since March. Long-term unemployment was little changed at 1.9 million people, or 27.1% of everyone out of work. Average hourly earnings rose 3.0% over the year to $37.81, and the average workweek stayed at 34.4 hours. 1
Two industries carry the detail worth keeping. Financial activities shed 7,000 jobs in September and is down 129,000 since a peak in May 2025, with insurance carriers and related activities accounting for 90,000 of that decline. Health care added 17,000, against a monthly average of 33,000 over the year, and nursing and residential care facilities lost 9,000. 1
For anyone weighing a move, the release describes a market that keeps the staff it has and adds few. The Bureau's own summary of the household survey says the unemployment rate "has remained in a narrow range" since March. 1
The AI-cited cuts kept arriving
HubSpot's board authorised its plan on 1 October and the company announced it on 6 October: about 660 positions, roughly 7% of the workforce, with $65 million to $75 million in severance and related charges falling mostly in the fourth quarter. Chief executive Yamini Rangan wrote that the shift to AI-driven customer outcomes is "transforming product, pricing and how we serve our customers", and that the company also needs "to fundamentally change the way we are organized". The Boston Globe reported that Rangan attributed the decision to the reorganization. 2
FICO's filing on 6 October covers about 15% of its staff. The company employed 3,811 people at the end of September 2025, which puts the reduction near 570 roles, and it expects about $27 million in pretax charges. The plan reduces management layers, simplifies the operating structure and integrates AI-driven product development. The reduction began in the week of the filing and is due to finish by the end of the third quarter of fiscal 2027. 3
Amazon confirmed on 6 and 7 October that it had eliminated fewer than 1,000 roles, mostly in the Stores business that runs its main shopping site. The cuts reached customer service, marketplace support and engineering, with staff in the United States, India and the United Kingdom affected, and the company said it had "adjusted parts of our Stores business" to deliver on its priorities. The reductions follow roughly 30,000 corporate roles cut from late last year into early this year, and they sit alongside the $220 billion the company plans to spend on AI infrastructure. In the same period Amazon has been contacting former employees about openings in AWS and AI roles. 4
The three filings describe three different relationships between AI and a job cut. FICO names AI inside the restructuring itself. Amazon removes retail roles while building AI capacity and hiring back into it. HubSpot's headline reads as an AI-led restructuring while its chief executive describes a change in organization design. A company copying a peer's move is copying one of those three, and the second half of the announcement usually says which.
The gains are landing on the workers who already had the better jobs
Gallup published the first results from the second year of the American Job Quality Study on 6 October, drawn from 15,482 employees surveyed between 26 January and 24 March 2026 for Jobs for the Future, The Families & Workers Fund and Gallup. 5
Among employees who use AI at work, 63% say it helps them finish tasks faster, 56% that it helps them find more creative solutions, 47% that it gives them more time on the parts of the job they find interesting, and 46% that the quality of their work improves. 5
Access is where the split opens. 28% of employees use AI at least weekly and 54% have never used it in their role. College graduates are more than twice as likely as workers without a degree to use it daily or weekly, at 40% against 17%, and managers use it at 37% against 25% among individual contributors. Roughly half of employees in professional and business services, financial activities, and information and media work with AI regularly, while more than 70% of healthcare support, manufacturing production and retail sales workers never do. Among employees who use AI weekly or more, 52% hold what the study defines as a quality job, against 46% of infrequent users and 32% of those who never use it. Gallup notes that the groups holding quality jobs and the groups using AI overlap, and that the direction of the relationship needs more research. 5
Loading chart…
Share of each group of U.S. employees who use AI at work daily or weekly, from the American Job Quality Study's survey of 15,482 employees, fielded 26 January to 24 March 2026. The education bars compare all employees by degree; the other two compare managers with individual contributors. Across the whole sample, 28% use AI at least weekly and 54% never do. 5
Front-line experience inside those groups differs as well. 31% of AI users say the technology has led their employer to ask them to take on more responsibility, and that share rises to 38% among users aged 18 to 24. 59% of regular users say AI lifts the quality of their work, against 26% of people who use it less often. Across the whole sample, 52% of employees say they have less influence over how new technology is adopted at their workplace than they would like. 5
A New York Fed study published on 8 October approached the same question from the employer's side. Working from the 2025 Small Business Credit Survey, Will Aarons and Asani Sarkar found that 46% of firms with at least one employee used AI, with another 15% planning to adopt it within a year. Firms already using it expected their headcount to rise by a net 33 percentage points, against 15 points among firms that did not. Adjusting for the owner's age, the firm's size, its profitability and its location narrowed the gap to 13 points, and adding the firm's own hiring over the previous year left 10 points. 6
The gap is widest where it is least expected. Among owners reporting fair or poor financial conditions, AI users' net expectation of adding staff ran 21 percentage points above non-users, against 10 points among owners in good shape. 63% of AI users said the technology was important to their production, 77% reported no change in their labour costs, and 31% reported higher sales. The study measures what owners expect over the next twelve months; its authors say whether those expectations turn into actual hires is a question for later data. 6
The obstacle is getting people into the growing roles
McKinsey Global Institute published its report Workforce in Motion: Skills and pathways to future jobs in the United States on 29 September, in the last days before this issue's window, and its numbers shaped much of this week's coverage. The report estimates that AI and automation will cut demand for the equivalent of about 36 million American jobs by 2035, while growth elsewhere creates about 41 million. In its base case, about 11 million workers, 7% of the workforce, would need to leave their occupation altogether, within a range of 6 million to 16 million. 7
The difficulty sits in the route between the two totals. McKinsey estimates that about 770,000 people a year would have to move between fields, roughly 3.6 times the historical average. One in seven displaced workers has a direct path into a growing job, which the report defines as one needing little retraining that pays at least as much. Nearly half face what it calls an "unpaved" path, blocked by skill gaps or credential requirements. About 85% of growing jobs require a credential, and 76% of them are done on site. The shrinking roles cluster in office and administrative support, retail and transportation; the growing ones in healthcare, construction and management. Lower-wage workers are 7.6 times as likely as higher-wage workers to need a new occupation. 7
Put that against the week's other numbers and the sequence is clear. The aggregate market is adding 29,000 jobs a month, so the external route is narrow. The workers who use AI most already hold the better jobs, so the internal route runs along lines the company drew long before AI arrived. And the roles that are growing need credentials and a place to work, both of which an employer can fund and neither of which a worker can supply alone. 157
What to do next
- Count who inside the company actually has AI, by group. The people lead should break AI use down by role and by business unit rather than reporting a single adoption percentage, because 54% of American employees have never used AI at work and access runs at twice the rate among graduates as among workers without a degree. A number that averages a manager and a warehouse floor tells nobody where the tool has landed. 5
- Name the roles the company will move people into, and who owns each move. The operations and people leads should write down which occupations the company expects to grow into, what credential or supervised experience each one needs, and who signs off on a transfer, since one in seven displaced workers has a direct path into a growing job and about 85% of those jobs need a credential. A funded certificate with a named destination is a plan; a general training budget reaches the people already closest to the tool. 7
- Read the whole announcement before citing a peer's AI layoff as evidence. The finance and people leads should note whether the filing names AI inside the restructuring, as FICO's does, whether the company is hiring back into AI roles, as Amazon is, or whether the chief executive describes organization design as the driver, as HubSpot's does. Those three cases support three different decisions. 234
- Weight the hiring plan toward site-based roles. Anyone setting next year's requisitions should note that 76% of the jobs McKinsey expects to grow are done on site, which puts them where the company's footprint, shift design and pay bands already sit, and that small firms using AI expect to add staff at a higher net rate than those that do not. 67
The week's readings point the same way. Payroll growth of 29,000 a month leaves almost no room to hire from outside. The AI-cited cuts are real and the reasons behind them vary. The workers who gain most from AI are already the ones holding the better jobs, and the roles that are growing need credentials and a place to stand. Every one of those facts points the next decision inward, to the company's own roles and training.
References
- 1
- 2
- 3FICO announces layoffs, management restructuring
housingwire.com
- 4
- 5AI Benefits at Work Unevenly Distributed
news.gallup.com
- 6AI Adoption and Employment Expectations: Evidence from a Survey of Small Business Owners
libertystreeteconomics.newyorkfed.org
- 7
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.