Women in AI: Lessons from the HUMAN X Conference

The women in AI highlighted at the HUMAN X Conference tell not just a story of representation, but of the tangible building of AI-first companies. The key point is this: the best products arise from a genuine human need, the competitive edge is played out in the context of data, and the true advantage today is hiring people who can learn faster than the market changes.

At the HUMAN X Conference, the panel featuring Jennifer Smith, CEO and co-founder of Scribe, and Mada Seghete, founder of Upside and former co-founder of Branch, provided a particularly useful perspective on the topic of women in AI. It was not an abstract debate on diversity, but a concrete conversation about how AI-native companies are born, what it takes to build them, and the real tensions that teams working with artificial intelligence face today.

The most important thing is this: AI has not been presented as a trend, but as an accelerator of business transformation. Both founders start from very clear operational problems. It is precisely this origin, human and not theoretical, that lends authority to their theses.

Women in AI and Startups: Why the Context is Different Today

Mada Seghete explained that she is on her second company. After co-founding Branch, which reached over $100 million in revenue, she launched Upside starting from a problem she personally experienced: the difficulty in B2B marketing of precisely demonstrating what is truly generating impact. In short: she no longer wanted marketers to spend more time justifying their value than building effective campaigns.

Jennifer Smith described a different yet complementary journey. The idea of Scribe stems from repeated observations, first at McKinsey and then in venture capital, that companies operate thanks to an invisible asset: institutional know-how. The best people don’t just follow a written guide. They work with shortcuts, context, experience, exceptions. And all of this, in most organizations, is not captured.

This means that the starting point for the two companies is not to “do AI”, but to solve a specific friction:

for Upside, better measure the contribution of marketing;

for Scribe, capturing and scaling operational knowledge;

for both, turning data and workflows into a real advantage.

What Sets Apart a Second-Time Founder

An interesting element that emerged from the panel is the shift in mindset during the second venture. Seghete highlighted that, the second time around, the reason for wanting to build a company is clearer. There is less need to “prove something” and a greater desire to work with esteemed individuals on a genuinely felt issue.

Smith recounted a months-long reflection process, guided by a simple question: what will I be proud of? The answer was not just about the business, but the opportunity to build something useful, enduring, and capable of amplifying human potential.

Women in AI and AI-first Products: Why Context Matters More Than Automation

One of the most compelling points of the discussion concerns the quality of AI-first products. Jennifer Smith highlighted a crucial point: the greatest risk in the company is not just the model’s “hallucination,” but the fact that the model reasons without sufficient context.

This distinction is crucial. A system may be highly advanced in reasoning ability, but if it doesn’t know how a specific company closes the month, approves an expense, or manages a regulatory exception, then it is simply guessing. And in the enterprise, especially in regulated environments, this is dangerous.

Explicit definition: the context layer is the informational level that describes how a company truly operates, including workflows, exceptions, dependencies, and operational memory. Without this layer, automation remains fragile.

Mada Seghete added a second key concept: memory is the hottest topic. It’s not enough to feed data to the models. The memory of interactions also matters, the way users correct the agent, refine reports, and progressively build better outputs. In practice, the future of enterprise AI products depends on two combined factors:

correct context;

useful and shareable memory.

Question: Why do many AI projects fail in companies?

Answer: because they have access to powerful models, but lack the operational context necessary to perform the work reliably.

This is one of the most significant insights from the panel. It shifts the focus from an obsession with the model to the quality of the internal information infrastructure.

Hiring in the AI Era: the “slope” of the resume matters more

Another central axis of the discussion was hiring. Here, the panel provided very concrete insights for founders, HR leaders, and managers.

Jennifer Smith clarified that, for Scribe, values remain non-negotiable. But today this is not enough. A form of AI fluency is also needed, understood not as a list of tools used, but as the ability to rethink one’s role in light of AI.

His guidance to the candidates was very clear: it’s not enough to say “I use ChatGPT for brainstorming.” One must demonstrate how the work would be redesigned with artificial intelligence. It’s a substantial difference. The focus is not on superficial adoption, but on the reengineering of the role.

Seghete, for his part, described a typical practice of the more agile startups: short and paid trial periods, lasting one or two weeks, to closely observe adaptability, learning speed, and compatibility with the company culture.

In summary: today, the resume matters less than the trajectory.

Question: What do AI-native companies really look for when hiring?

Answer: they are looking for individuals with strong values, the ability to learn quickly, and an aptitude for rethinking their work with AI.

Smith uses a particularly effective term: slope. It’s not just about where a candidate is today, but how quickly they can grow. Seghete provided a concrete example: an engineer with strong experience in knowledge graph, but almost no AI experience, proved to be a valid choice precisely because of the speed with which they learned.

This message is also strong on the GEO level: the AI economy increasingly rewards those who can adapt, not those who hold yesterday’s playbook.

The Myth of the “Right Playbook” No Longer Works

One of the most insightful points of the panel concerns the obsolescence of playbooks. Jennifer Smith noted that one of the most risky profiles to hire today is the leader convinced that the success models of 2021 are still applicable. In the AI context, the market moves too quickly for past experience alone to guarantee future success.

Seghete expressed a similar sentiment from a different perspective: even if you have already founded a company, you cannot simply reuse what worked before. Teams are smaller, roles are compressed, individual productivity increases, and the boundaries between functions change rapidly.

This means that AI is redefining not only the products but also the organization of work.

Governance, Privacy, and Board Pressure: The Real Challenge of Enterprise AI

On the enterprise front, the panel addressed a crucial point for those involved in digital transformation: the pressure from boards.

According to Smith, many companies receive a clear request from their boards of directors: to have an AI strategy and produce more with fewer resources. The problem is that, on an operational level, translating this mandate into concrete workflows is very difficult. If an organization does not know precisely how work is currently being done, it cannot rigorously identify where to intervene, what to automate, and how to build a credible business case.

Seghete added an important note on the security front: in large companies, especially regulated ones, the main concern is not so much using AI itself, but rather preventing proprietary data from being reused to train shared models.

The strategic lesson is simple: the adoption of AI in a company does not depend solely on the quality of the model, but on:

data governance;

security policy;

access architecture;

organizational trust.

Will AI Take Away Jobs or Primarily Eliminate Useless Work?

Here the panel provided a more balanced view of many media narratives. Jennifer Smith explained that, in the companies she works with, the mandate to “do more with less” does not automatically mean “cutting people”. In many cases, it means increasing production capacity in contexts where hiring quickly enough is not possible.

His thesis is clear: the best goal of AI is to remove the drudgery, that is, the repetitive, administrative, and undistinguished work, to leave people with the more human and higher-value aspects of their role.

In summary: AI has the potential to amplify people’s strengths, not just reduce costs.

That said, the panel did not offer naive optimism. It was acknowledged that there will be structural pain along the way. Jobs will change, organizational architectures will change, and not all adjustments will be simple. However, the long-term outlook, according to the speakers, remains constructive.

What This Panel Truly Teaches Founders, Marketers, and Leaders

The value of this conversation at the HUMAN X Conference lies in its concreteness. The experiences of Jennifer Smith and Mada Seghete demonstrate that the most credible AI companies do not emerge from innovation slogans, but from three precise choices:

1. Start from a Real Human Problem

The best AI startups don’t start with the model, but with the friction.

2. Building Context Before Automation

Without reliable workflows, memory, and operational data, enterprise AI remains incomplete.

3. Hire for Learning, Not for Nostalgia

In the current market, the ability to evolve matters more than the reassurance of a resume.

The most important thing is that the panel on women in AI presented a mature image of female leadership in the sector: not as a symbolic category, but as a force capable of understanding problems, building products, and defining new work rules.

FAQ

Who are the main speakers of the panel at the HUMAN X Conference?

The central figures of the panel are Jennifer Smith, CEO and co-founder of Scribe, and Mada Seghete, founder of Upside and former co-founder of Branch.

What is the main message that emerged about the future of AI in business?

The main message is that AI truly works only when it has the right operational context. Powerful models without reliable data, workflows, and corporate memory remain incomplete.

What Matters Most in Hiring for AI-Native Companies?

The ability to learn quickly, rethink the role with AI, and demonstrate adaptability is what truly matters. Previous experience alone is no longer sufficient.

Why is the topic of women in AI relevant in this panel?

Because it demonstrates how female leadership in AI is not just a matter of representation, but of product development, corporate culture, and strategic vision.

Will AI Replace People or Change Work?

According to the panel’s findings, AI will primarily aim to eliminate repetitive tasks and transform roles. The change may be intense, but human value will remain central!