McKinsey estimates that about 770,000 U.S. workers a year may need to change occupations as AI and automation spread, and that women are 1.6 times as likely as men to be among them.

History repeats itself. Back in the 1920s, many women were employed as switchboard operators , until AT&T automated the process, so callers could connect directly with each other. And of course, today that job has largely disappeared.

So, what happened to those women operators who lost their jobs? After almost a decade , the former operators were no longer working or were in lower-paying jobs.

Since generative AI tools like ChatGPT, Gemini and Claude arrived, ILO and others have identified administrative jobs as the most exposed. Women are heavily represented in them: ILO finds women’s jobs are nearly three times as likely as men’s to fall in the highest exposure category in high-income countries, driven by clerical roles.

Similar clerical jobs grew to absorb young women after telephone operator jobs disappeared. Among workers aged 22 to 25, a Stanford Digital Economy Lab and ADP report finds 43.8% of women work in the most exposed occupations versus 32.4% of men, largely because of the occupations women hold.

The Broken Rung of admin: The automation of these jobs affects both the women who hold them today and the women seeking entry-level jobs. Furthermore, this worsens the broken rung problem that McKinsey and LeanIn identified, because fewer entry-level roles means fewer first promotions to manager.

In their 2025 report , only 93 women were promoted to manager for every 100 men, and women hold just 29% of C-suite roles. At entry level, 33% of men are encouraged by their manager to use AI tools versus 21% of women.

Shift to New Work (and the barriers)

New roles are being created, growing most in healthcare, construction, and management, but more than 75% of the workers who must switch come from office support, retail, and transportation, and the path is harder for women.

1. A McKinsey report states that only 14% of transitions are a direct transfer of skill sets, and men are about 30% more likely than women to have those direct pathways.

2. 85% of these new jobs require credentialing.

3. 76% of these jobs cannot be done remotely.

A good example in the McKinsey report is the move from an office assistant to project manager. The office assistant has 58% of the skills required and could see a 154% increase in pay. But, here is the catch: Project management typically demands a bachelor’s degree (which only 47% of assistants hold) and credentials like a PMP or CAPM, requiring 3–5 years of relevant experience.

Where do we go from here?

Companies and governments can start now with three steps.

1. Track women’s pathways. Set KPIs for the promotion rate to manager and the AI encouragement gap, and course correct as soon as the numbers slip.

2. Hire and train for skills. Many corporates are leaning on upskilling, with 77% of employers planning it, yet few transitions are direct, tracked via KPIs like direct-fit mobility. Currently only ~14% of workforce transitions occur between closely adjacent roles without substantial reskilling, so reskilling matters more. Only about 38% of growing jobs have a legally required credential, so employers control much of the barrier, and skills-based hiring can lower it.

3. Fix child care. Government can take steps to solve the child care situation that often makes in-office jobs a huge challenge and raises the cost of workforce participation. With 76% of growing jobs unable to be done remotely, child care will shape who can take them. Caregiving , including child care costs, was the strongest reason women gave for leaving the workforce voluntarily in 2025. Government action is uneven: The 2025 tax law raised the dependent care account limit to $7,500 , a change the Tax Policy Center calls modest.

I have spent 25 years inside large companies, including as an executive at Alphabet/Google, Tesco, and Ford. In my experience, budgets for upskilling are significant relative to reskilling, which reflects a trade-off between operational continuity and workforce transformation. Today, the need is workforce transformation.

Almost a century ago, the cost of automating telephone switching fell mostly on women. Let us not let history repeat itself. We can be smarter.