Women at work · Two Deloitte reports · Version 1, November 2026
Deloitte asked 7,500 women what helps and blocks them at work. Separately, it asked 3,500 senior leaders how their companies are using AI. Read side by side, the two reports raise questions neither one answers alone.
The two reports
Deloitte Global · 2026
7,500 women in 15 countries, surveyed May to June 2026: progression, flexibility, safety, care, health, and why women stay or leave.
Read the report →Deloitte · October 2026
3,500 director to C-suite leaders in 25 countries and six industries: how companies use AI, redesign jobs, and govern it. The words “women” and “gender” do not appear in it.
Read the report →Where the reports meet
Women @ Work 2026: 43% of women surveyed want to leave their current job. Of those, 45% name lack of career progression as a main reason, and 42% the chance to earn more doing the same job elsewhere.
State of AI in the Enterprise: Only 15% of companies report a significant or fundamental shift in decision-making authority as a result of AI. Deloitte names roles, authority, accountability and career paths as what still needs redesigning (page 19).
Read together: as companies redesign jobs around AI, who receives the new decision authority and the new career paths? Are women in the roles where that authority is being created?
Options for her (choices, not tasks):
What an employer could do:
Who owns it: the executive sponsoring the redesign, with human resources.
How to check it worked: compare who was eligible, who was invited and who was selected for redesigned roles, before and after the change.
Limit: a before-and-after difference does not by itself prove the change caused it.
Women @ Work 2026: 80% believe flexible working hurts career progression at their company, and 85% believe their workload would not be adjusted if they reduced their hours.
State of AI in the Enterprise: 81% say AI has made decisions and execution faster, but only 25% are using it to transform the business deeply (page 6). When companies add AI but keep the old approval steps, Deloitte warns they can create more work rather than more capacity (page 19).
Read together: when AI makes work faster, does the time saved become flexibility, or more work for the same people?
Options for her (choices, not tasks):
What an employer could do:
Who owns it: the manager of the team using the AI tool.
How to check it worked: compare hours worked and workload with flexible-work uptake and promotions, before and after.
Limit: Deloitte’s speed figure is reported by leaders. It does not measure anyone’s workload.
Women @ Work 2026: Only 54% believe they can report non-inclusive behaviours at work without any career impact. Among women who did report, satisfaction with their employer’s response ranged from 9% to 29%, depending on the behaviour.
State of AI in the Enterprise: When an AI system produces a bad outcome, 36% of companies hold the team that owns it responsible, 24% a central governance function, and 10% the employee who prompted the system. 23% say accountability is unclear, so problems can move between teams with no one obliged to resolve them (page 21).
Read together: where blame lands on the person who prompted the system, who are those people? And if women already doubt that reporting is safe, will they report an AI error at all?
Options for her (choices, not tasks):
What an employer could do:
Who owns it: the governance lead for AI, or the team owner where there is none.
How to check it worked: count error reports, the time taken to close them, and whether people who reported say they would report again.
Limit: Women @ Work asked about non-inclusive behaviours, not AI errors. The link is a question to test, not a finding.
Women @ Work 2026: 58% worry about learning the skills they need for future job security, and 57% are concerned that AI may replace their job.
State of AI in the Enterprise: For physical AI, such as robotics, Deloitte finds adoption is rising faster than expertise (page 8). It describes one retailer that kept its designers and repurposed and upskilled them rather than cutting them (page 19).
Read together: if expertise lags adoption, who is chosen to build it? Does training reach the women who fear for their jobs?
Options for her (choices, not tasks):
What an employer could do:
Who owns it: the learning and development lead, with each manager.
How to check it worked: compare who was eligible, invited, took part and later got AI-related work, broken down by gender.
Limit: the expertise finding is about physical AI only, and the retailer is one example Deloitte describes, not a measured result.
Women @ Work 2026: 53% say the combination of their job and home responsibilities is not manageable, and 29% say they feel burned out.
State of AI in the Enterprise: 40% of companies let AI make decisions or take actions without human approval to at least a moderate degree, and Deloitte finds the systems to oversee those decisions are lagging behind (pages 8 and 20).
Read together: when AI acts first and people review afterward, who does that review, and is the time counted in their workload?
Options for her (choices, not tasks):
What an employer could do:
Who owns it: the governance lead for AI, with the manager of the reviewing team.
How to check it worked: compare review time logged with review time assigned, and track error rates.
Limit: the 40% is a share of companies, not of decisions. Nothing in either report links it to burnout. This raises a question; it shows no connection.
Where they don’t
These findings matter on their own. The AI report adds no evidence about them, so no comparison is forced. A research question appears only where it is useful, and it is labelled as a question, not a finding.
Women @ Work 2026: Only 14% say their career has progressed as they wanted since returning from leave, and 39% say it took time to regain their confidence.
What State of AI does not tell us: Nothing about leave, returning to work or parents.
A question worth researching: When a role is redesigned around AI while someone is on leave, what is she returning to?
Women @ Work 2026: 45% are concerned about at least one aspect of personal safety connected to work, including 23% at work and 20% on the commute.
What State of AI does not tell us: Nothing about personal safety. Its physical AI section is about machinery and operations, not the safety of the people working there.
Women @ Work 2026: Only 27% of women with children under 18 have access to affordable childcare. 43% of them have had to take unexpected time off, and 31% of those gave their employer a different reason.
What State of AI does not tell us: Nothing about care.
Women @ Work 2026: Among women living with a partner and children under 18, 51% hold primary responsibility for childcare, compared with 5% for their partner.
What State of AI does not tell us: Nothing about the home.
Women @ Work 2026: 23% have experienced health challenges related to menstruation, perimenopause, menopause or fertility. Some of these groups are small in the survey.
What State of AI does not tell us: Nothing about health.
Women @ Work 2026: Career progression, flexible working, a safe and inclusive culture, mental health support and work/life balance are the strongest predictors of how long women expect to stay.
What State of AI does not tell us: No link to these conditions. Its call for clear decision rights, escalation paths and accountability for AI (page 21) is related in spirit, but it is not evidence about them.
A question worth researching: When a company builds decision rights and escalation paths for AI, do the same routes carry women’s concerns about these five conditions?
What the comparison shows
Women say what holds them back: progression, workload, unsafe reporting and fear for their jobs. Leaders say AI is making work faster, but that authority, accountability and expertise are lagging behind. Neither survey asks whether those gaps fall harder on women. That is the question worth testing: with one woman’s account, the pattern across many women, and a change an employer makes and then checks.
Independent analysis by Darrell Ellens. Not commissioned, reviewed or endorsed by Deloitte. All figures are Deloitte’s, describe the people surveyed, and are linked to the original reports; Women @ Work figures are from the report’s key findings and chapters, State of AI page numbers are shown. This page is updated as the analysis is reviewed. If you see an error, please email darrell@darrellellens.com.
If you lead an organization redesigning work around AI, or you support women going through it, I’d be glad to talk.