Key Insights
Operationalising AI Requires More Than Just Models
Justin Bundick emphasizes that deploying AI at scale within a large enterprise is not just about building smart models—it’s about having the right infrastructure, data governance, and organisational alignment. At Southwest Airlines, AI solutions like the Leading Indicators Alert Tool (LIAT) were only possible because of the foundational work put into their cloud-native Integrated Data Foundation and AI Common Platform. These platforms enable experimentation, deployment, and real-time integration into business workflows. Justin outlines how successful scaling requires thinking beyond POCs and investing in resilient systems, continuous model monitoring, and clean, well-labelled data. For technology executives, the takeaway is clear: you cannot shortcut infrastructure readiness if you want AI to deliver sustained business value.
Designing AI with Empathy Enhances Adoption and Impact
For Justin, AI is most valuable when it augments—not replaces—human decision-making. He shares how his team embeds empathy into AI design by immersing themselves in the business context before building any solution. Whether through shadowing frontline operations or mapping the actual friction points in workflows, this human-centric approach ensures the AI outputs are both relevant and usable. Justin stresses that effective AI must mirror the decision patterns of real users and present insights in digestible, action-oriented ways. By aligning closely with how people actually work, AI becomes an enabler rather than a disruptor. This insight is especially important for executives seeking adoption at scale: solutions must not only be technically sound but also designed for real-world integration.
Strategic Partnerships Accelerate AI Maturity
Rather than building every component in-house, Southwest Airlines leverages strategic partnerships with AWS and Anthropic to scale its AI capabilities. Justin explains how AWS infrastructure underpins their AI practice, enabling secure, production-grade deployments of both traditional and generative AI models. Through Amazon Bedrock, they’re also able to test and apply Anthropic’s pre-trained transformer models, giving them speed and flexibility without sacrificing quality. These partnerships have helped the airline stay agile in a rapidly evolving space, enabling internal teams to focus on high-value, enterprise-specific use cases. For executives navigating limited internal AI talent or time-to-market pressure, Justin’s approach illustrates how external partnerships can be a force multiplier—when paired with clear business objectives and a well-defined technical architecture.
Episode Highlights
AI’s Role in Managing Operational Complexity
Justin explains how Southwest Airlines built the Leading Indicators Alert Tool (LIAT) to proactively manage disruption across its complex network of aircraft, crew, customer, cargo, and maintenance operations. He describes how the convergence of these networks creates a high-risk environment and how predictive AI tools provide critical visibility and control in real time.
“We experimented with the possibility of being able to build an alerting system that allowed us to identify where we might have degradation in the network itself. And so that’s where the Leading Indicators Alert System came into fruition.”
From Business Immersion to Better AI Design
Justin shares how empathy-driven design begins with deep engagement in business processes. His team spends time with frontline operations to fully understand the needs, behaviours, and decisions that AI is meant to support. This ensures AI outputs are not just technically sound, but contextually useful and trusted by their end users.
“We analyze the data on the back end to try to understand the patterns inside of the data and how that might match what we actually saw in real life. And by doing that, it helps us create AI that’s more human-centric and more aligned to the human need.”
Using Generative AI to Accelerate Ideation
Justin discusses how generative AI is helping internal teams push their thinking during the early phases of product design. From defining features to imagining adoption scenarios, GenAI supports the creative process and helps uncover new possibilities faster than traditional methods.
“I’m seeing some really powerful approaches to using generative AI in those early define and design phases of the process to really help push our thinking about how a product should be designed, what should be included in that product, and how it could be adopted by the users.”
The Hidden Barriers to Scaling Enterprise AI
Justin outlines a pragmatic view of what holds most enterprises back from scaling AI—from misaligned infrastructure and siloed data to talent scarcity and change resistance. He emphasizes the importance of starting foundational work early, even while exploring use cases.
“You need to start that work now, even while you’re working on your infrastructure and your data and your talent pipelines, to help set the tone and the narrative of how your company and enterprise is going to utilize AI—and hopefully utilize it as an augmenter of people.”
