
Building an AI Center of Excellence: Lessons from Iron Mountain's Journey
Why We Created Our AI Center of Excellence (CoE)
When ChatGPT burst onto the scene at the end of 2022, it was uncharted waters for many. Everybody was trying to figure out what their AI strategy should be. At Iron Mountain, we acknowledged that we don't know what we don't know, and we needed a diverse team of experts to help us understand and implement AI—not just technology experts, but legal experts, compliance experts, and people who know how the work gets done in our business units.
The Four Pillars of Our AI CoE
- Cross-Functional Collaboration We wanted to put together a core cross-functional group to educate the collective “us” on what AI is, how legal teams are looking at it, what it means from a compliance standpoint, and what the practical applications are. Getting the teams together was one of the main things we did as part of establishing the AI Center of Excellence.
- Demystifying AI Through Education The CoE is responsible for demystifying AI. A lot of people think AI is out there to take people's jobs. We don't look at it that way. We see this as not a job displacement tool, but rather as a job augmentation tool—another tool in our toolkit. We're thoughtful about educating the entire Iron Mountain community on what AI is, what its limitations are, what its applications are, and enabling and empowering them to come up with ideas while incentivizing them to implement those ideas.
- Establishing Security and Compliance Foundations We wanted to bring AI in securely and compliantly. We established our non-negotiables around security, compliance, data privacy, data residency, and preserving entitlements as new data gets generated through the system. These are foundational and non-negotiable for our company. We plugged this into our Architecture Review Board, where we review third-party risk and have legal review of any new products we bring in. This foundational work was a huge part of establishing those necessary components.
- Measuring Return on Investment We are not deploying AI for the sake of deploying AI. AI is not a hammer looking for a nail. We need to ask: How do we responsibly, securely, and compliantly embed AI into a customer's journey through the Iron Mountain ecosystem?
We understand what processes underpin that customer journey and embed AI into relevant, pertinent areas in secure, compliant, responsible, and ethical ways. Once we embed it, we track the outcomes: cost optimization, better customer satisfaction, improved time to value, or new products and capabilities.
The Results Speak for Themselves
Our approach has been wildly successful. The Massachusetts Technology Leadership Council (MTLC) recently awarded Iron Mountain’s InSight Digital Experience Platform (DXP) its Tech Top 50 Award in AI Implementation. We have around 1,200 to 1,300 people taking these trainings daily, coming up with ideas. We get together as a committee, review them, and support the execution of those activities.
The Incentive Structure
We incentivize participation. If you come up with an idea, you get points. If you implement the idea, you get bigger points. This could mean company swag from the company store or spot bonuses.
Initially, for the first three months, we came up with brilliant ideas that didn't get picked up—nobody wanted to use them. That's when we realized we needed to change our approach to focus on implementation, not just ideation.
Looking Ahead: The Future of AI Centers of Excellence
Will Iron Mountain still have an AI Center of Excellence in a decade? I would be lying if I told you I knew the answer to that. What I do envision is that as we embark on this journey of agentic AI and agent-to-agent frameworks, we're going to have tons of agents and digital workers deployed in the field.
But the foundational elements—security, compliance, data privacy—are going to look totally different 10 years from now. There's always going to be that element that requires some semblance of a cross-functional group thinking about governance and compliance more so than the technology bits.
So I do believe the CoE would exist in a slightly different format than it is right now, focused more on governance structures than on technology democratization.
This blog post is based on an original AIIM OnAir podcast. When recording podcasts, AIIM uses AI-enabled transcription in Zoom. We then use that transcription as part of a prompt with Claude Pro, Anthropic’s AI assistant. AIIM staff (aka humans) then edit the output from Claude for accuracy, completeness, and tone. In this way, we use AI to increase the accessibility of our podcast and extend the value of great content.
About Swami Jayaraman
Swami Jayaraman is Senior Vice President of Global Technology and Chief Enterprise Architect at Iron Mountain, where he leads the company's Artificial Intelligence Center of Excellence. With over 20 years of technology leadership experience, he specializes in enterprise architecture, cloud computing, and AI transformation across highly regulated industries. Before joining Iron Mountain in 2019, Swami spent eight years as Senior Vice President at Bank of America, where he managed complex technological ecosystems in the financial services sector. His unique background spans both financial services and information management, giving him deep insights into how organizations can leverage technology to manage, protect, and extract value from their data assets. As a recognized thought leader in enterprise architecture and artificial intelligence, Swami serves on advisory boards for Nutanix and HMG Strategy, helping to shape industry innovation and direction. At Iron Mountain, he drives AI initiatives that transform how businesses approach information governance and digital transformation in an increasingly data-driven world.