When AI Reshuffles the Board, Culture Decides Who Remains at the Table
Europe is not running out of women in technology; it is running out of women in roles that define the future.
The critical figure is not just the drop to 19% in core roles but the organizational design that pushes talent toward functions with less power and, increasingly, greater risk of automation.
A report from McKinsey published in early 2026, based on the analysis of 4 million LinkedIn profiles and additional data, estimates that women represented just 19% of jobs in "core" technology roles in 2025, a 3 percentage point decline from the previous year. In a similar snapshot, the gap widens at higher levels: 13% in management roles and only 8% in senior management (director and C-level positions). The primary reason for this exit is not “lack of supply” but work culture.
Source
This is not just a discussion of corporate sensitivity. It is a matter of business architecture. When a company expels part of its talent from positions that build products, data, and infrastructure, it is making a portfolio decision, even if it does not acknowledge it: it protects the present but narrows the future.
The Uncomfortable Truth: It’s Not Total Participation but the Core That Generates Advantage
Here lies the mirage that confuses many executive committees. On one hand, Ravio reports that women make up 40% of the total tech workforce in Europe, with relatively stable figures across countries (for example, 43% in Spain and 39% in the Netherlands). However, that average includes roles that, while critical, do not always control the technological direction or platform budget. In parallel, that same report places women at 21% in executive roles across Europe, with lows of 17% in Spain and Sweden.
Source
The contrast with McKinsey is understood through operational definitions: broad “tech workforce” versus “core tech roles” (engineering, software, AI, data, infrastructure). When the core drops to 19%, the company is losing diversity precisely where the product architecture, data model quality, and delivery speed are determined. This impacts competitiveness, not reputation.
Moreover, segmentation by roles reveals a pattern of “risk concentration”: McKinsey reports that women are overrepresented in design (54%) and product management (39%), but these functions are a smaller portion of the total and have a lower automatic conversion to P&L and C-level positions in many companies. If the pipeline to direction passes through technical ownership and platform roles, the statistics do not lie.
Source
From my portfolio lens, this is equivalent to investing the majority of female human capital in "layers" of the stack with less control over technology strategy, and then being surprised when leadership does not balance out. It’s not a mystery. It's design.
The Exodus Accelerates Where Companies Treat Culture as an Indirect Cost
The report is explicit: work culture is the main driver behind the exit of women from core roles in Europe.
Source
In corporate terms, “culture” often degrades to a set of soft initiatives or an annual survey. This approach fails because real culture is an operating system: who decides, who receives visible projects, how performance is evaluated, what work is rewarded, and what work is rendered invisible.
McKinsey provides an operational indicator that many CFOs would immediately understand: women spend on average 200 hours a year on “office housework”, administrative or coordinating tasks that sustain the organization but rarely count as career progression.
Source
That is not an anecdote. It is a systematic redirection of capacity from value-creating activities (build) to internal maintenance activities (run). If an organization recurrently assigns this "glue work" to a group, it is also recurrently assigning less time to deep technical learning, fewer commits, less exposure to critical systems, and diminished impact narratives.
In companies operating under margin pressure, bureaucracy finds a perfect justification: standardizing, committee-making, adding layers. The result is that the most talented individuals with options leave first. And in technology, the talent most likely to have options often resides in core roles.
The reputational risk exists, but the financial risk is more immediate: turnover in core roles raises recruitment and retraining costs and, above all, reduces delivery speed just when AI forces a reshuffling of the roadmap.
AI Does Not 'Create' the Gap but Punishes Organizations that Ignore It
The most underestimated dynamic in the report is the change in demand across levels and functions. From 2024 to 2025, McKinsey observes sharp declines in entry-level roles for women in product development (−17%) and software engineering (−13%). At the same time, the roles of AI, data, and analytics grow for women by 7%, but for men by 11%.
Source
This produces a scissor effect. The organization reduces entry points in functions where technical muscle is built and seniority is constructed while accelerating demand in areas where a larger male base already exists. If a culture that repels is added to the mix, the gap amplifies itself.
There is a portfolio design reading behind this: many companies are utilizing AI to optimize the present (automate, reduce layers, gain efficiency) before using it to explore the future (new products, new architectures, new data models). That sequence is not inherently incorrect, but it has collateral costs: cuts tend to fall on entry-level and "hybrid" roles that lack strong sponsorship within engineering.
The consequence is that female talent, already more concentrated in certain functions, faces a market with fewer ramps to access growth tracks. Without ramps, there is no mobility. Without mobility, there is no retention.
The report also marks a structural contrast: in the European Union, women already make up 40.5% of scientists and engineers (7.9 million in 2024, up from 3.4 million in 2008), but this critical mass does not translate into participation in corporate core tech.
Source
The pipeline exists; conversion fails.
What the C-level Typically Mismeasures and What Should Govern Without Bureaucracy
When a company claims it “cannot find women” for core roles, it is usually describing its internal funnel. McKinsey notes that 33% of bachelor’s degrees and 39% of PhDs in tech are awarded to women, but the system breaks down thereafter: from career choice to insertion and retention in core roles.
Source
My experience shows that this type of exodus is rarely corrected by more policies and almost never through internal campaigns. It is corrected when talent governance is integrated into the governance of technological portfolios. Three operational fronts, without romanticism:
1) Metrics that distinguish “production” from “maintenance”. If there is an annual stock of 200 hours of unrecognized work, the solution is not to ask for “more collaboration.” It is to instrument: rotational assignment, explicit recognition, and load limits by role. What is not measured is assigned by inertia.
2) Real mobility toward areas that are growing. If AI, data, and analytics are growing more for men than for women, the problem is not just external supply. It is about transition mechanisms: reskilling with sponsorship, bridge projects, and evaluation by deliverables. McKinsey suggests AI-assisted reskilling as an “on-ramp” in response to the contraction of entry roles.
Source
That point is crucial if executed as an internal product with accountability, not as an optional course.
3) Representation with management cadence, not as an annual snapshot. McKinsey mentions quarterly objectives for representation and promotion based on output.
Source
Quarterly matters because it forces small corrections and avoids the problem being disguised in next year’s budget.
The enemy here is human resources bureaucracy disconnected from the business. The natural owner of this agenda is the leadership of technology and product alongside finance, as it relates to future installed capacity. The organization that delegates this as an “issue of culture” ends up paying with speed and innovation.
Competitive Advantage Will Belong to Those Who Protect the Present’s Revenues Without Shrinking the Future
Europe is entering a reconfiguration of tech work where AI displaces entry-level tasks and elevates the value of roles with technical judgment, solid data, and the ability to industrialize models. In that context, the drop to 19% in core roles and the bottleneck of 8% in senior management do not describe an employer brand issue but a weakness in organizational design.
Source
Companies that want to compete do not need a manifesto. They need to protect the current revenue engine while financing exploration with real autonomy, and they need female talent to participate in that exploration from the core, not from peripheral functions. The viability of the model is defined by the ability to maintain profitability today without sacrificing access to the talent that builds the products and infrastructure of tomorrow.
