We recently gathered an exceptional group of leaders, founders, engineers and advocates for our Women in Tech: Agentic AI for Inclusion, Diversity & Impact event in Sydney in partnership with Luxury Escapes. The discussion centred on one of the most critical opportunities in technology today: how Agentic AI can reduce barriers, expand access and create more inclusive workplaces.

What followed was a practical, honest and deeply forward-looking conversation with over 80 senior technology leaders in the room. Our panel included: 

• Laetitia Andrac, Co-Founder & CEO of Understanding Zoe and neurodivergent advocate, is reshaping how AI supports neurodivergent individuals and communities.
• Dominick Ng, Director of Engineering at Relevance AI, is leading teams building the world's leading AI Workforce platform.
• Ishita Gupta, Founding Engineer at Kinso, blending AI cognition, robotics and policy to shape next-generation AI.
• Nathan Scully, Co-Founder & CTO of Adora, is building agent-powered product analytics and documentation tools.
• Moderated by Jody Weir, future-of-work design leader with deep expertise in Agile, Lean and transformational ways of working.
• Hosted by Marie Higgins, Principal at Allura Partners, specialising in Software Engineering recruitment and championing diversity through her fast-growing Women in Tech community.

Moving from Reactive to Proactive Systems

The panel opened by sharing how they use Agentic AI within their organisations. For Laetitia, the technology is foundational, not optional. She described her team's AI-first approach: they built their platform around "Pip, the AI at the core of our platform," an autonomous agent that works with an internal evaluator agent that "scores its responses, provides feedback, and alerts us in real time when a response is inappropriate"

This set the stage for defining Agentic AI itself. As Laetitia explained, it represents a shift from passive assistance to goal-seeking autonomy: "AI systems that have agency… can act with a degree of autonomy toward a goal, not just respond to prompts."

Dominick framed it through a problem-solving lens, describing Agentic AI as the ability to tackle "goal-oriented vs. deterministic problems."

Together, their definitions reinforced a shared understanding: Agentic AI isn't about better prompts, it's about AI that can reason, plan, remember and act.

Confronting Bias in Today's Models

Bias in AI was a central theme, and Laetitia did not shy away from the reality: "Most AI today is still built on neurotypical assumptions, which makes it quietly ableist."

She pointed to research showing major models often associate terms like autism and ADHD with danger or disorder, a direct reflection of the datasets they are trained on. This is why her team intentionally built the opposite:

"Our AI is trained and audited by neurodivergent individuals and experts to use strength-based, neuroaffirming language… It replaces deficit terms like 'low functioning' and ‘high functioning’ with 'high support needs' and ‘ low support needs’"

Ishita emphasised that organisations must proactively shape their data foundations, noting that to challenge systemic bias, AI must "include data sets regarding inclusion… A chatbot would need to include data sets on local concerns."

The message was clear: inclusive AI is designed, not assumed.

Guardrails for a High-Autonomy Technology

With autonomy comes risk, and the panel acknowledged that missteps can be significant. Referring to well-known industry incidents, they highlighted the need for structured safeguards, including evaluator agents, real-time monitoring, and consistent human oversight.

Laetitia emphasised that continuous review becomes essential precisely because Agentic AI evolves: its behaviour changes as it learns, which means vigilance cannot be optional.

Levelling the Playing Field with Agentic AI

A powerful thread emerged around how Agentic AI is reshaping career mobility, especially for early-career professionals and underrepresented groups.  Nathan highlighted that tools with such leverage can enable people to rise faster than ever before, explaining that when new technology creates this kind of force multiplier, individuals can "leapfrog progress."

Ishita expanded on this, describing a moment of unprecedented opportunity: "People are now in demand with AI skills; it is a level playing field to get started."

Both agreed that the path forward is hands-on experimentation: find a problem and build an agent to solve it.

Practical Steps for Getting Started

The panel offered clear entry points for different backgrounds:

For non-technical professionals:

• Start with no-code tools like Relevance AI or n8n
• Automate small tasks to build intuition
• Learn through building, not theory

For engineers:

• Explore agent frameworks and cognitive architectures
• Experiment with tool-calling, memory and planning loops
• Study how agents behave in real environments

Across both paths, the advice was consistent: practice is the real teacher. Dominick outlined where organisations are already seeing value, noting that AI is unlocking growth "without explosive hires," and that companies are beginning to stand up dedicated AI Ops teams focused on identifying the best Agentic AI opportunities across the business.

The Future is Agentic, and Inclusive

The discussion made one thing clear: Agentic AI is more than a technological trend; it’s a tool for transformation. When intentionally designed with inclusive data, strong safeguards, and diverse perspectives at the helm, it can break down barriers, accelerate careers, and reshape workplaces for the better.

From neuroaffirming language models to career-leapfrogging agents, the potential is real, and the call to action is urgent. Organisations and individuals who experiment boldly, centre inclusion, and maintain rigorous oversight won’t just keep pace; they will define the future of work.

Agentic AI is here to act, and so are the leaders shaping it. The question now isn’t if we leverage its potential, it’s how intentionally we do so.