Citi’s Strategic Approach to AI Adoption
While many large organizations limit artificial intelligence initiatives to small, isolated projects, Citi has taken a markedly different approach by integrating AI into the daily workflows of approximately 4,000 employees. Over the past two years, the bank has expanded AI use beyond specialist teams to a broad cross-section of its workforce, spanning technology, operations, risk management, and customer support roles.
Building an Internal AI Workforce
Citi’s internal AI initiative began with cultivating a network of “AI Champions” and “AI Accelerators”—employees who volunteer to learn about AI tools and help their colleagues adopt them. This grassroots strategy focuses on empowering individuals within teams rather than relying solely on centralized control. These champions receive training, access to approved AI systems, and internal resources, enabling them to act as local advisors rather than formal instructors.
By distributing AI knowledge throughout the organization, Citi has fostered a peer-driven adoption model that encourages practical and responsible AI use. Employees earn internal badges by completing courses or demonstrating AI applications in their work, which helps build credibility and visibility without tying directly to promotions or salary increases.
Widespread Use of AI Tools in Daily Tasks
More than 70% of Citi’s global workforce of roughly 182,000 employees now use firm-approved AI tools to improve efficiency. These tools assist in summarizing documents, drafting internal communications, analyzing data sets, and supporting software development. The emphasis is on augmenting routine tasks to reduce friction and enhance productivity, rather than pursuing radical innovation.
Importantly, Citi enforces strict governance by limiting AI use to approved platforms with clear guidelines on data handling and output management. This cautious approach aligns with the bank’s regulatory obligations and helps build managerial confidence in expanding AI access across departments.
Lessons for Large Enterprises Scaling AI
Citi’s experience highlights that successful AI integration at scale hinges on training enough employees to use the technology responsibly and confidently, rather than turning everyone into experts. This decentralization reduces dependence on a small team of specialists and integrates AI more naturally into everyday work.
The bank’s approach also challenges the notion that AI adoption must be top-down. While leadership endorsement was critical, much of the momentum came from motivated employees volunteering to learn and teach AI tools, underscoring the importance of bottom-up engagement in technology adoption.
However, peer-led adoption requires ongoing support to maintain enthusiasm and consistency. Citi addresses this by rotating AI Champions and updating training materials to keep pace with evolving AI capabilities.
AI as Infrastructure, Not Just Innovation
Citi treats AI as a foundational tool embedded in existing processes rather than a disruptive technology seeking to transform the business overnight. This mindset allows for measurable progress through incremental improvements and reduces the pressure on dramatic outcomes.
As many companies struggle to move AI projects from pilot phases to full production, Citi’s model provides a valuable example of how distributing ownership across teams, combined with centralized governance, can accelerate adoption and build lasting capabilities.
In summary, Citi’s quiet yet methodical rollout illustrates how empowering employees and embedding AI into daily work can drive meaningful change in large organizations, offering a replicable blueprint for others navigating the complex AI landscape.
Fonte: ver artigo original

GPT-5.2 Brings Practical Enhancements to AI Language Capabilities
Meta Expands Renewable Energy Portfolio with Additional 650 MW Solar Capacity to Support AI Growth
ChatGPT-Driven Referrals to Retailer Apps Surge 28% Year-Over-Year on Black Friday
Global AI Show Abu Dhabi 2025 Highlights Innovation and Leadership in Artificial Intelligence