Customs and Border Protection is using artificial intelligence coding tools to deliver capabilities to its Automated Commercial Environment faster.
AI Tools in Action at CBP Customs and Border Protection (CBP) has long utilized AI for tasks like facial recognition. Now, AI is being applied to accelerate IT modernization, specifically in converting legacy code to modern components, leading to a 40% reduction in estimated timelines. Thomas Mills, from CBP, emphasized that while AI-driven code generation is powerful, human oversight remains crucial to review the output. AI is also used for automated data extraction and data governance to ensure the quality of data fed into large language models (LLMs), preventing 'garbage in, garbage out' scenarios in their Automated Commercial Environment (ACE) system modernization efforts. Iterative development to predict costs CBP is implementing an iterative development approach, using two-week sprints to deliver new capabilities to users. This methodology helps the agency to better estimate the holistic costs of fielding capabilities incrementally, especially concerning cloud expenses. By breaking down development into small, manageable chunks, CBP can assess outcomes quickly, learn from failures, and scale successful initiatives more effectively. This systematic approach supports informed funding advocacy based on tangible use case findings. AI for Mission and Back Office Functions at Commerce The Department of Commerce is leveraging AI to modernize IT and enhance both mission-critical and administrative functions across its bureaus. Matt Passos, Commerce's acting CTO, highlighted the primary goal of making their 'dumb data smart' by using AI. They are partnering with GSA and utilizing the USAi platform, alongside APIs, to give older systems access to AI's processing power. This helps mission experts clean and standardize data, which is essential before applying AI to improve functions. AI is proving particularly effective for capital planning, helping to determine total cost of ownership, project success/failure, and gamify investment decisions. Reducing the risk of AI vendor lock in Agencies are concerned about the cost and potential vendor lock-in associated with AI tools. Matt Passos noted the uncertainty in future AI pricing and the challenge of quantitatively demonstrating AI's value. To address this, GSA's USAi platform aims to provide a shared service that mitigates vendor lock-in. Ryan Palmer, USAi's architecture lead, explained that the platform offers access to various AI models from different providers, allowing agencies to confidently explore, test, and integrate solutions without being tied to a single vendor. This approach helps agencies to 'rent' the AI model they need, promoting flexible modernization and data unlocking capabilities.
AI Tools in Action at CBP
Customs and Border Protection (CBP) has long utilized AI for tasks like facial recognition. Now, AI is being applied to accelerate IT modernization, specifically in converting legacy code to modern components, leading to a 40% reduction in estimated timelines. Thomas Mills, from CBP, emphasized that while AI-driven code generation is powerful, human oversight remains crucial to review the output. AI is also used for automated data extraction and data governance to ensure the quality of data fed into large language models (LLMs), preventing 'garbage in, garbage out' scenarios in their Automated Commercial Environment (ACE) system modernization efforts.
Iterative development to predict costs
CBP is implementing an iterative development approach, using two-week sprints to deliver new capabilities to users. This methodology helps the agency to better estimate the holistic costs of fielding capabilities incrementally, especially concerning cloud expenses. By breaking down development into small, manageable chunks, CBP can assess outcomes quickly, learn from failures, and scale successful initiatives more effectively. This systematic approach supports informed funding advocacy based on tangible use case findings.
AI for Mission and Back Office Functions at Commerce
The Department of Commerce is leveraging AI to modernize IT and enhance both mission-critical and administrative functions across its bureaus. Matt Passos, Commerce's acting CTO, highlighted the primary goal of making their 'dumb data smart' by using AI. They are partnering with GSA and utilizing the USAi platform, alongside APIs, to give older systems access to AI's processing power. This helps mission experts clean and standardize data, which is essential before applying AI to improve functions. AI is proving particularly effective for capital planning, helping to determine total cost of ownership, project success/failure, and gamify investment decisions.
Reducing the risk of AI vendor lock in
Agencies are concerned about the cost and potential vendor lock-in associated with AI tools. Matt Passos noted the uncertainty in future AI pricing and the challenge of quantitatively demonstrating AI's value. To address this, GSA's USAi platform aims to provide a shared service that mitigates vendor lock-in. Ryan Palmer, USAi's architecture lead, explained that the platform offers access to various AI models from different providers, allowing agencies to confidently explore, test, and integrate solutions without being tied to a single vendor. This approach helps agencies to 'rent' the AI model they need, promoting flexible modernization and data unlocking capabilities.
