Data and AI-driven decision-making are reshaping industries, but what does it take to successfully implement and scale these solutions? In two insightful 2024 Everyday AI panel discussions, Luc Gagnon, director of analytics, travel marketing, & loyalty spend at Air Canada, and Vincent Huang, director of data science, chief audit executive group, at Royal Bank of Canada (RBC), shared their journeys of transforming their organizations through AI, balancing innovation with governance, and driving measurable business impact.
From optimizing marketing campaigns to building next best action decision engines, these leaders highlighted key challenges, strategies, and outcomes that can inspire any organization looking to operationalize AI at scale.
Let’s dive into the most impactful lessons from these discussions.
Building AI for Real Business Impact
One of the strongest themes across both panels was the emphasis on solving real business problems rather than chasing AI for AI’s sake.
We may have the best models, the best data scientists, the best tools, but if they’re not used and they remain on the shelf, then they’re useless. It’s all about working closely with the business.
— Luc Gagnon, director of analytics, travel marketing, & loyalty spend at Air Canada
At RBC, the internal audit team took this approach by automating risk assessment using AI models that proactively flag potential risks before they escalate. This transformation moved their process from reactive to proactive, reducing manual work and improving decision-making.
Automating for Efficiency and Scale
For many organizations, AI’s biggest value lies in automation — freeing up teams to focus on strategic work while eliminating manual, repetitive processes.
At Air Canada, the marketing analytics team faced a backlog of post-campaign analyses that took weeks to complete. Instead of continuing with a slow, manual process, they stepped back and rebuilt the workflow with Dataiku, reducing the analysis time from two weeks to just five-and-a-half hours.
We needed to stop the car, change the square wheels for round ones, and suddenly we could go much faster and much further.
— Luc Gagnon, director of analytics, travel marketing, & loyalty spend at Air Canada
Balancing AI Innovation With Governance
As organizations scale AI, governance and compliance become critical concerns. AI models must be explainable, auditable, and compliant with industry regulations.
At RBC, strict regulatory requirements meant that AI models needed to be fully transparent. To address this, the team built a structured risk assessment tool (RAPTOR) that provides objective risk scores across departments, helping internal auditors make data-driven decisions.
Governance is also crucial for ensuring data integrity. Air Canada tackled this by implementing centralized data platforms and monitoring tools to improve data quality and consistency across teams.
Democratizing AI: Empowering Business Users
One major challenge in AI adoption is scaling knowledge beyond technical teams. Both Air Canada and RBC emphasized giving business teams the tools to interact with AI models — without needing deep technical skills.
At Air Canada, marketing teams now use self-service AI applications to generate customer insights without waiting on data scientists.
Before, only data scientists could run our recommender system. Now, marketers can trigger campaigns themselves, freeing up data scientists to focus on more advanced work.
— Luc Gagnon, director of analytics, travel marketing, & loyalty spend at Air Canada
For RBC, empowering non-technical teams with AI helped reduce bottlenecks and accelerate adoption across the company.
If everything is urgent, then nothing is really urgent. By putting AI into business users' hands, we reduce fire drills and let data teams focus on high-value projects.
— Vincent Huang, director of data science, chief audit executive group, at RBC
What’s Next? The Future of AI Innovation
All three companies are already looking ahead to the next wave of AI advancements, including LLMs, causal inference, and deeper automation.
At Air Canada, the next challenge is automating causal inference models at scale to measure not just correlation, but causation.
We can already detect causality of events (as opposed to correlation), but understanding causality at scale is our next big project.
— Luc Gagnon, director of analytics, travel marketing, & loyalty spend at Air Canada
For RBC, the focus is on enhancing AI-driven risk detection to improve proactive risk mitigation strategies.
We’ve moved from reactive to proactive. Now, we want to go even further — anticipating risk before it even surfaces.
— Vincent Huang, director of data science, chief audit executive group, at RBC
Final Thoughts
Across these two panels, the biggest AI success factors were:
- Align AI With Business Goals: AI (including GenAI) is only valuable when it solves real business problems.
- Automate for Scale: AI should eliminate manual work and accelerate decision-making.
- Governance is Critical: Compliance, transparency, and explainability are non-negotiable.
- Empower Business Teams: AI adoption grows when non-technical users can engage with AI models.
- Future-Proof AI Strategies: The next step for AI is causal inference, LLM-powered automation, and deeper personalization.
These organizations aren’t just adopting AI — they’re redefining how their industries operate with data-driven intelligence.