After five years at Westpac, scaling our data team from 20 to over 1,000 people, I’ve learned that implementing AI at scale is less about the technology and more about the people, processes, and culture you build around it.
The Reality Check
Everyone wants to talk about AI strategy. C-suites are excited. Boards are asking questions. Consultants are pitching frameworks. But here’s what I’ve learned from the trenches: most companies aren’t ready for AI, and that’s okay.
The problem isn’t the technology—it’s rarely the technology. The problem is that AI amplifies everything about your organisation: your data quality, your processes, your culture, your decision-making speed. If those things are broken, AI will just help you fail faster.
What Actually Works
Let me share three things that made the difference for us at Westpac, and that I’m applying now at Cochlear:
1. Start with the Problem, Not the Solution
I can’t count how many times I’ve been in meetings where someone says “We need to use AI for [insert business function].” That’s backwards. The question should be: “What’s the biggest pain point for our customers or business?” Then, and only then, ask if AI is the right solution.
At Westpac, we didn’t set out to build AI models. We set out to reduce customer complaints, speed up loan approvals, and detect fraud earlier. AI was the tool, not the goal.
2. Build the Foundation First
You can’t have AI without data. And you can’t have good data without:
- Clear data ownership and governance
- Infrastructure that can handle scale
- People who understand both the business and the tech
- A culture that values data-driven decision making
This isn’t sexy work. It’s unglamorous infrastructure building, policy writing, and culture change. But it’s essential. I tell CEOs: “Give me 6 months to fix your data foundation, or your AI initiative will fail within 18 months.”
3. Hire for Curiosity, Not Just Credentials
When we scaled from 20 to 1,000 people, we made plenty of hiring mistakes. The biggest lesson? The best data scientists and AI engineers aren’t always the ones with PhDs from Stanford.
The best ones are curious. They ask why. They want to understand the business. They’re comfortable with ambiguity. They can explain complex concepts to non-technical stakeholders.
I’d rather hire someone with mid-tier technical skills and high curiosity than a technical genius who can’t communicate or understand the business context.
Looking Forward
The future of enterprise AI isn’t about fancy algorithms or bigger models. It’s about organizations that can:
- Make decisions faster with data
- Adapt quickly to changing conditions
- Combine human judgment with machine intelligence
- Build trust with customers about how AI is used
At Cochlear, we’re applying these lessons to healthcare AI—a space with even higher stakes. The principles remain the same: start with the problem, build the foundation, hire curious people, and focus on business value, not just cool technology.
Final Thoughts
If you’re embarking on an AI journey at your company, my advice is simple: slow down. Take the time to understand your current state. Fix your data foundation. Build a team that understands both business and technology. Create a culture of experimentation and learning.
The companies that win with AI won’t be the ones who move fastest—they’ll be the ones who build the strongest foundations and maintain the patience to do it right.