Scaling AI inside a large organization isn’t just about technology — it’s about people, processes, and strategy. At Everyday AI Boston, John Irvin, Chief Data Officer for Deloitte Technology, shared an inside look at Deloitte’s AI journey with Dataiku, detailing how they democratized data science and tackled real-world challenges.
His session wasn’t about theory or high-level platitudes — it was a candid, practical discussion on what it really takes to scale AI inside a $35 billion company with 180,000 employees.
Building an AI-Driven Organization
John emphasized that AI adoption isn’t just about technology — it’s also about trust and ethics. Organizations scaling AI need to ensure responsible and transparent practices, which is why Deloitte has made ethics and governance a priority.
He pointed to Deloitte’s Data AI Institute as a resource for staying informed on the latest AI trends, research, and specialist perspectives. He also highlighted Deloitte’s Trustworthy AI framework, which provides guidance on ethical AI practices, risk management, and governance. As generative AI continues to evolve, this framework is being updated to reflect new challenges and best practices.
If you’re not talking about ethics and trust with AI, you should be.
- John Irvin, Chief Data Officer for Deloitte Technology, Deloitte Consulting LLP
But responsible AI is only part of the equation — scaling AI also requires the right tools and processes. That’s where Deloitte faced its next challenge.
From Fragmented Data Science to a Unified Approach
Nine years ago, Deloitte’s internal data science teams were scattered across the organization, each using their own tools and workflows. There was no standardization, no common process, and no streamlined way to scale insights across the organization.
To address this, Deloitte launched the Deloitte Data Science Lab, a centralized platform designed to give internal data scientists easier access to data, automation, and analytics tools. It was a massive improvement — but it didn’t solve everything. The real challenge came when Deloitte realized AI wasn’t just for experienced data scientists. There was an entire ecosystem of analysts, engineers, and business teams who also needed AI-driven insights.
Hence, five years ago, Deloitte started searching for a platform that could serve both experienced data scientists and non-technical users. John gave his team one simple directive: “You don’t have to come back with one tool. But if you do, it better be the right one.”
And they did. Dataiku was the unanimous choice.
- For experienced data scientists, it helped make their work faster and more efficient.
- For analysts and business users, it provided an intuitive low-code/no-code environment.
- For Deloitte, this decision unlocked a milestone: AI adoption at scale, beyond just technical teams.
Scaling AI Adoption: The Power of Blended Teams
Even with the right tools, Deloitte faced another challenge: People. With 180,000 employees but a limited number of experienced data scientists, demand quickly outpaced supply. So Deloitte tried something different: blended teams.
A case study from Deloitte’s Government Solutions & Innovation team highlighted this approach. With only two experienced data scientists on a 20-person team, demand was overwhelming. The solution? Pairing each specialist with three to five citizen data scientists.
One breakthrough came from redefining AI’s role in workforce management. Traditionally, Deloitte classified professionals based on compliance-driven requirements (skills, experience, billing rates). But they flipped the script — using AI to identify skill gaps, accelerate employee growth, and even unlock new business opportunities.
The result? More efficient workforce planning, stronger insights, and a better employee experience (while driving revenue growth).
The Toughest Challenge: Resistance From Within
One of the most notable challenges encountered was the initial hesitation among some data scientists and traditional analysts. These data scientists were accustomed to working independently and needed some time to adjust to the blended team model, which emphasizes collaboration with non-technical users. Similarly, traditional analysts required support to embrace AI technologies. The term “citizen data scientist” itself created a mental barrier, making some analysts feel like they weren’t part of the AI conversation.
To overcome this, Deloitte launched a Lighthouse Use Case Program, collaborating with Dataiku to showcase real-world success stories. Teams were invited to submit use cases, receive hands-on training, and present their results. The only requirement? They had to form a blended team.
This approach worked. By demonstrating tangible wins and encouraging peer-to-peer knowledge sharing, Deloitte shifted internal attitudes and proved that AI wasn’t just for specialists — it was for everyone.
Key Lessons From Deloitte’s Journey to Scaling AI
John’s session was more than just a recap of Deloitte’s AI strategy — it was a real, unfiltered look at what it takes to scale AI inside a massive organization.
His key takeaway? AI success isn’t about building more models — it’s about creating an environment where AI can thrive. And for Deloitte, that meant:
- The right tools
- The right structure
- The right mindset shift
Because the biggest AI challenge isn’t technology — it’s people. And once you solve that, everything else falls into place.