Have you decided on a distinct operating model for AI? Is it different than IT or D&A? Is this centralized, distributed, federated?
Sort by:
This is a common question for us. Our discussions with Gartner experts revealed that AI’s placement varies by company, depending on existing structures and evolution. Some organizations combine data and AI, while others keep them separate. Currently, we keep them separate. Initially, AI was placed under innovation to ramp up capabilities, but it’s clear that in a few years, AI will be embedded in all IT departments. For now, governance is managed separately, with CBAs created for all AI programs and weighted by value. We avoid pursuing every opportunity, focusing on those with clear returns and ensuring business and IT alignment. Our implementations span homegrown solutions, embedded AI in existing systems, and new programs like Excelia and Extracto. Experiences vary, especially with customer-facing AI agents. Governance is collaborative, but long-term success will depend on strong, trusted data and metadata, which ties back to the data department. We continue to evaluate where AI should reside organizationally.
Our approach is similar to Aleks’. As a regulated entity subject to SOX and SEC regulations, governance and compliance are critical. We adopted a centralized model, and I developed our AI strategy for the organization and affiliates, grounded in governance, compliance, and industry alignment. I participate in an asset management industry group, sitting on its AI committee, which helps us stay aligned with industry trends across various firm sizes. Our strategy includes an AI advisory component within our governance framework and a cross-functional team to ensure broad understanding. We execute initiatives in three core areas: AI and automation, generative AI, and AI agents. We are careful not to overwhelm the organization with too many tools, given the rapid release of new solutions. Managing cost and adoption is crucial, and the greatest lesson has been that AI success requires a top-down initiative. AI fundamentally changes operations, and isolated pilots are unlikely to succeed unless tightly contained.
Currently, we favor a centralized approach for all intake and deployment of AI tools. Due to strict compliance, certifications, and governance requirements in our data center environment, security remains a top priority. This drives our preference for centralized deployment, though we do incorporate cross-functional requirements and requests. Our method involves a cross-functional AI working group led by an IT leader, with regular participation from business unit leaders to discuss AI needs and business cases. We focus on value, outcomes, cost, and delivered benefits. Additionally, we have an AI steering committee at the leadership level that handles approvals for new solutions, tools, activities, and funding. This committee also conducts deeper reviews into security and governance. Since our acquisition by American Tower three years ago, we have established an AR steering committee between both entities to further collaboration. We are progressing on this journey,
Similar to thoughts shared. Two nuances, we structured governance in an AI Transformation Management Office that has a truly cross-functional team to help weigh in on and evaluate policy and updates. This has helped us move faster than our prior efforts which were led more siloed in S&T and legal. Lastly, we have started to give Federated responsibility to leaders within each of the sub-businesses so they can take on more of their local responsibilities and use cases. Getting time from those federated leaders is challenging but will hopefullly help us move quickly enough to support their growing and shifting needs.