Leadership is entering a new chapter. As AI becomes embedded in everyday decision-making, the question is no longer whether organizations will adopt AI, but how leaders will guide its use with judgment, accountability, and trust.
The technology is advancing quickly, but leadership is being tested in more fundamental ways. According to the World Economic Forum's Future of Jobs Report 2025, analytical thinking, resilience, flexibility, leadership, and social influence are among the skills expected to grow most in importance over the coming years. At the same time, women continue to represent only around 25–30% of the global AI workforce, while fewer than 15% of executive AI leadership positions are held by women. As AI becomes increasingly embedded in the systems shaping healthcare, finance, cybersecurity, education, and public services, the perspectives of the leaders guiding its development have never mattered more.
The most effective leaders are no longer distinguished by how much they know, but by how they think, how they make decisions when certainty is impossible, and how they balance technological capability with human responsibility.
To explore what this new era demands, WomenTech Network spoke with four Executive Women in Tech (EWIT) members whose experience spans AI strategy, cybersecurity, healthcare, operations, data, and emerging technologies.
Alexandra Sfrijan
COO and Co-Founder of CYBALGORIS
Ewa J. Kleczyk
Chief Data & Analytics Officer at Prolaio
Gaytri Khandelwal
Global VP, AI Sales & Partnerships at Crest Data
Vidya Shankaran
Field CTO, Emerging Technologies at Commvault
Leading With Curiosity and Clarity
Technological change has made certainty an increasingly unrealistic standard for leadership. Decisions that once remained useful for years can lose relevance within months, while traditional planning cycles struggle to keep pace with new capabilities and changing expectations.
The response is not to chase every development or abandon long-term direction. It is to create an organization that can learn and adjust without losing sight of its priorities.
“Leaders can no longer rely on having all the answers. The pace of change requires leaders to create environments where learning, experimentation, and adaptation happen continuously.”
That shift also changes where authority comes from. Vidya Shankaran describes it as a move “from command-and-control to context-and-clarity.” When a technology decision that worked last quarter may be obsolete the next, adaptability becomes more useful than relying on tenure or past experience alone. Tomorrow’s strongest leaders, she says, will “ask the sharpest questions, decide with incomplete data, and unlearn faster than the market changes.”
Ewa J. Kleczyk approaches the same change from the relationship between leaders and technology. Leadership must evolve from directing people alone to guiding “a mixed team of people and systems,” with clear judgment about where a human must reinforce the connection between them. That responsibility requires leaders to understand how a model works well enough to interrogate it, without assuming they can out-compute it.
Moving with change should not mean acting impulsively. Gaytri Khandelwal begins by asking, “What’s the actual upside if this goes well, and what’s the real damage if it doesn’t?” When the value clearly outweighs the risk, a phased introduction allows leaders to observe what is happening and adjust before early problems become larger ones.
Keeping people informed is central to that process. As Gaytri Khandelwal explains, the difference comes from “having a real plan around it—keeping people in the loop, watching how it’s actually playing out, and being willing to adjust instead of stubbornly sticking to the original plan.”
Leadership in this environment is less about predicting every change and more about helping people respond to it with confidence. Teams need enough clarity to move, enough room to learn, and enough trust to reconsider a plan when the evidence no longer supports it.
How Leadership Is Evolving
As organizations rely more heavily on data, automation, and intelligent systems, the qualities associated with strong leadership are changing. Technical fluency matters, but it is only part of what leaders now need.
Curiosity, adaptability, and critical thinking sit at the center of Alexandra Sfrijan's view. Leaders must understand both the opportunities and limitations of AI while maintaining their teams’ confidence that technology is being used responsibly and that business decisions remain aligned with human values.
Gaytri Khandelwal makes a similar distinction. “We leaders don’t need to be data scientists,” she says, “but we must deeply understand how data and automation impact our business, systems, and people.” Without that foundation, it becomes difficult to know when an automated recommendation deserves confidence and when it needs further scrutiny.
Ewa J. Kleczyk identifies three additional qualities: intellectual humility, structured skepticism, and translation. Intellectual humility allows leaders to acknowledge what neither they nor a model know. Translation enables them to explain what a model is doing to regulators, clinicians, boards, and other nontechnical stakeholders in language that supports informed judgment.
“The second is structured skepticism: treating automated output as a hypothesis to be tested rather than a verdict to be accepted, until it has been checked against ground truth.”
In regulated industries, translation is no longer simply a communication skill. It supports audit-readiness, safety, adoption, and ultimately the impact a system can have.
Vidya Shankaran describes another essential quality as “judgment under uncertainty”—knowing “when to trust the model, when to override it, and when to pause.” She pairs that judgment with “healthy paranoia,” the discipline to ask “what could break this?” before focusing only on what a system makes possible.
Ethical courage completes that picture. Leaders must sometimes be willing to slow a deployment when governance is not ready, even when organizational pressure favors speed.
Empathy remains equally important. As machines assume more routine decisions, Gaytri Khandelwal expects “the human parts of leadership—motivating people, navigating conflict, reading a room” to become the real differentiator. Technology may inform a decision, but leaders still have to understand how that decision will be experienced by the people around them.
These qualities do not compete with technical understanding; they make that understanding useful. AI leadership depends on knowing enough to question the system, remaining humble enough to recognize uncertainty, and retaining the human awareness needed to judge what should happen next.
Leading With Confidence Through Change
Some leadership practices are becoming less useful as information becomes more accessible and intelligent systems change how work is organized. Moving beyond them does not mean dismissing experience or structure. It means recognizing where they have become too rigid for current conditions.
“The assumption that expertise alone creates authority is becoming less relevant,” Alexandra Sfrijan says. Technology can provide insights at a scale no individual leader can match, while useful knowledge is increasingly distributed across employees, data systems, and intelligent tools. Collaboration, transparency, and continuous learning become more valuable when hierarchy alone cannot provide the full picture.
Gaytri Khandelwal sees command-and-control leadership aging out for a similar reason: “Nobody can keep up with everything anymore.” A leader who can say, “I don’t know, let’s figure it out together,” may create more value than one who projects certainty unsupported by evidence.
Planning practices also need to become more responsive. Annual cycles are increasingly insufficient when AI can change workflows within months. As Gaytri Khandelwal puts it, “The ground shifts monthly.” Quarterly and yearly planning can provide direction while preserving room to respond to new information.
“Rigid roadmaps can’t survive AI-era velocity; we need rolling, evidence-led pivots.”
Evidence should also carry more influence than status. Vidya Shankaran questions the HiPPO model—the Highest-Paid Person’s Opinion—in which seniority outweighs data and frontline knowledge. People closest to the work may see emerging problems earlier, and leadership must make room for those signals.
The meaning of productivity is changing as well. Gaytri Khandelwal challenges “the old habit of rewarding busyness and long hours over actual output” when automation can absorb repetitive work. As intelligent assistants become more widely available, some previously siloed responsibilities may merge, making outcomes more meaningful than visible activity.
Speed requires similar reconsideration. Ewa J. Kleczyk notes that “‘faster’ served as the leadership scoreboard” for years, but speed without a human checkpoint can scale bias and error rather than catch them. In domains affecting human health, review gates are not barriers to progress; they are part of making that progress reliable.
The practices being left behind share a common weakness: they place too much confidence in fixed authority, fixed plans, or speed alone. Leading confidently through change requires something more durable—clear direction combined with broader input, shorter feedback cycles, and a willingness to revise decisions when circumstances change.
The New Strength of Adaptable Leadership
Human judgment, creativity, and connection do not become less important as AI systems gain greater autonomy. They become more important because leaders must decide where autonomy belongs, where intervention is necessary, and who remains answerable for the outcome.
Human-in-the-loop oversight cannot be reduced to placing a person somewhere in an automated process. A reviewer who lacks the time, information, or authority to challenge a recommendation may technically remain involved while contributing little meaningful judgment.
Ewa J. Kleczyk therefore treats human oversight “as a design principle rather than a compliance checkbox.” In healthcare, the most valuable checkpoint is where a domain expert can interrogate a model’s reasoning—not merely approve its final output—before it informs a patient-facing decision.
The boundary should move only when evidence supports it. Leaders must continually define “where autonomy ends, and human judgment begins,” Ewa J. Kleczyk explains, adjusting that boundary only as reliability accumulates.
“The key is treating AI as a collaborator that handles scale and speed, while humans stay in charge of purpose, ethics, and meaning—the ‘why’ behind the work.”
Keeping teams close to customers and real human experiences helps preserve that “why.” AI can amplify people’s creativity and judgment, but it should not displace the moments where those qualities matter most.
“AI should augment human capabilities, not replace human responsibility,” Alexandra Sfrijan says. Even when execution is delegated to a system, leaders remain accountable for outcomes and for keeping human judgment central to decisions affecting people. The opportunity is to improve creativity, productivity, and problem-solving while preserving empathy, ethics, and meaningful human relationships.
Vidya Shankaran translates that responsibility into practical controls. “Agents act, humans answer,” she says. Every autonomous action should connect to a human-owned policy, access boundary, and audit trail. Leaders should also “build pause, override, and rollback into agentic workflows—not as afterthoughts but as first-class controls.”
AI may optimize execution, but people must retain ownership of intent, ethics, and customer trust. Human-centered innovation does not require limiting every form of autonomy. It requires clear boundaries around that autonomy and identifiable people who remain able—and required—to answer for its consequences.
Better Questions, Stronger Leadership
AI systems are beginning to influence decisions in healthcare, employment, credit, safety, and public life. Representation among the people building and governing them therefore affects more than workforce demographics. It influences which problems are prioritized, which risks are identified, and whose experiences inform the design.
“The systems we’re building today will mediate healthcare, hiring, credit, and safety for generations,” Vidya Shankaran says. “If the people designing them aren’t representative, the blind spots get encoded at scale.”
Direct participation matters. Gaytri Khandelwal argues that women need “a literal seat at the table” in technical roles, funding decisions, and governance bodies because influence often belongs to those “actually building and deciding, not just advising.”
Vidya Shankaran makes that participation more specific: women should claim positions at AI architecture, model-governance, and red-team tables. Publishing and presenting can help women shape the wider discussion, while sponsoring others extends that influence to the next generation. As she puts it, “Mentorship is force multiplication.”
Visible role models and trusted communities make those routes more accessible. WomenTech Network and Women AI Builders provide places where women can exchange ideas, ask questions without being judged, develop professional relationships, and see clearer pathways into technical and executive decision-making.
“Representation in AI is not solely a matter of fairness; it changes what gets built and who is positioned to catch blind spots before they become embedded bias.”
Alexandra Sfrijan believes representation must extend beyond technology roles into the positions where AI strategies, policies, and governance frameworks are defined. “Diverse perspectives lead to better outcomes, stronger innovation, and more responsible technology,” she says. Professional communities help women influence the future of technology rather than simply adapt to decisions made elsewhere.
Ewa J. Kleczyk's own guiding principle is to “create your own table and convince others to join.” Rather than waiting indefinitely for a seat, women can build the room and make it worth others’ time to join them there.
Individual achievement matters, but professional communities can turn that achievement into collective influence. Vidya Shankaran describes their value as “visibility, credentialing, and the networks that turn capability into authority.” Her conclusion makes the broader governance case plainly: “Representation isn’t a diversity metric; it’s a risk control.”
Four Executive Perspectives
Before wrapping up the conversation, we invited each executive to consider one question shaped by their own work, experience, and area of leadership. Their responses bring a more focused perspective to the themes explored above, showing how the broader demands of AI leadership are taking shape across different professional contexts.
How can leaders embed security and responsibility into innovation from the beginning, especially as AI becomes part of more products and systems?
“Security and responsibility cannot be treated as final checkpoints. They must be part of the design process from the start. Building secure and trustworthy systems requires proactive governance, continuous risk assessment and a culture where innovation and responsibility evolve together rather than compete with one another.”
COO and Co-Founder of CYBALGORIS
How can data and AI help leaders make better decisions while keeping innovation responsible, ethical, and human-centered?
"Data and AI meaningfully expand what leaders can see patterns across far more information than any individual could hold in mind, but responsible use requires that leaders remain the ones asking why a model produced a given output, not only what it produced. That habit distinguishes AI that earns durable trust in healthcare from AI that quietly erodes it."
Chief Data & Analytics Officer at Prolaio
How can leaders connect AI innovation with strategy, partnerships, customer value, and real-world market impact?
"In strategy, allowing for marketing automation and simulations before spending real money. For partnerships, it's about picking the right allies by comparing a multitude of dimensions. For customer value, AI can be deployed as support to detect and solve problems before the customer escalates. For market impact, use cases include smarter pricing, leaner supply chains, and spotting opportunities."
Global VP, AI Sales & Partnerships at Crest Data
As AI becomes part of critical systems and cyber resilience strategies, what should leaders understand about trust, risk, and readiness?
"As AI moves into the critical path, leaders must internalize three truths: Trust is earned at the data layer. Risk has shifted from perimeter to behavior. Readiness is a continuous discipline, not a checklist. The leadership mandate is simple: if you can't trust it, control it, or recover it — you can't scale it. Resilience over panic. Readiness over reaction."
Field CTO, Emerging Technologies at Commvault
Leadership Is Still a Human Choice
Technology will continue to evolve, and leadership will evolve with it. AI can help organizations see more, move faster, and approach complex problems in new ways, but its value will ultimately depend on the decisions leaders make: where autonomy belongs, when human judgment must intervene, whose perspectives are included, and who remains accountable for the outcome.
The opportunity is not simply to adopt more capable systems. It is to build organizations where curiosity, responsibility, trust, and human understanding develop alongside them.
Continue the conversation at the AI Builders Global Conference 2026 (October 14-15, Virtual), nominate an inspiring woman shaping technology for the Women in Tech & AI Global Awards, and become part of the WomenTech Network community advancing more inclusive and responsible leadership in technology.