Accelerating Development by 40% Using Generative AI and Amazon Q Developer with DTCC

DTCC provides clearing, settlement, transaction processing, data tools, and trade reporting services to the global financial services community. The company wanted to use generative AI to further enhance its innovation initiatives, adding business value and creating new features for clients. However, market infrastructure providers like DTCC are held to high regulatory standards, so the use of emerging technologies such as generative AI requires careful consideration and thorough testing. Accordingly, DTCC crafted a unified generative AI adoption program to implement generative AI responsibly and effectively throughout its business units. “We wanted to improve developer productivity and get the right tools in the right hands to democratize the power of generative AI across DTCC,” says Johnna Powell, managing director and head of technology, research, and innovation at DTCC.

The company began using Amazon Q Developer to assist software engineers as they wrote and refactored code. That use case demonstrated the capabilities of the service and proved its potential business value to the organization’s leadership. DTCC completed a proof of concept (POC) in April 2024. “Using Amazon Q Developer throughout the organization, we could aid workflows and elevate our overall developer productivity,” says Behar Cami, managing director of technology at DTCC.

After running the POC for 4 months in 2024, DTCC determined that it was ready for production. “We were really pleased with the performance,” says Cami. “We surveyed our software engineers, and they said it enhanced their productivity and their ability to do their jobs better. Our POC teams enjoyed using the tool and highlighted that it improved their experience.” Encouraged by this success, DTCC leadership approved a wider release for Amazon Q Developer.

With the POC complete, DTCC is rolling out Amazon Q Developer to production in phases, opening it up to several hundred software engineers at a time. Meanwhile, it will use Amazon Bedrock, a fully managed service that offers a choice of high-performing large language models, to expand its AI rollout to other use cases. The company’s generative AI adoption is driving change across the organization while enhancing productivity, accelerating development, and simplifying workloads for engineers.

Team members can use Amazon Q Developer for assistance with code generation, testing, and debugging. They can also get help with business questions, their code base, and other topics related to enterprise data. Using the service, DTCC engineers can transform and implement new code to meet their specific requests. During the POC, the company boosted developer throughput by 40 percent, accelerating product development by saving many hours a week. “During our POCs, we found that something that used to take 10 hours now takes 6 hours using Amazon Q Developer,” says Powell. “That is quite powerful.” The company also reduced code defects and the associated operational risks by 30 percent while improving test coverage during the POC. Throughout the use of Amazon Q Developer, build-failure rates did not increase.

DTCC currently has 10–15 generative AI–based solutions in its POC, including internal and client-facing tools. In one use case, DTCC is using Amazon Q Developer to power an internal, searchable database so that employees can derive actionable insights and deliver increased value and levels of service. In addition, the company is improving its application security, risk mitigation, and overall efficiency with Amazon Q Developer. Through its use cases, DTCC has improved security scores across its code repositories by 5 percent. “We’re driving generative AI democratization across DTCC, and we want to do it in a safe way,” says Powell. With security built into AWS services, DTCC can protect its software engineers’ work and clients’ data in the cloud under the AWS shared responsibility model. “We have a framework that we use to score various tools and providers, and AWS stood out in terms of our robust requirements for compliance and risk management,” says Powell.

DTCC is also proactively mitigating risks by publishing a generative AI policy that provides upfront guidance and establishes a governance structure within the organization. “It was very important for us to establish a centralized location where we keep a repository of all the generative AI initiatives going on across DTCC for two reasons: First, we wanted to avoid duplication of efforts and increase efficiencies, and second, we wanted to use it for inventory and tracking purposes,” says Powell. The company established a team of 40 people who are responsible for monitoring generative AI initiatives and tracking approvals for in-progress POCs. That group includes members from the company’s technology risk management team, model risk management team, and IT security architecture team, as well as other stakeholders, and reports monthly on generative AI projects.

DTCC aims to expand access to its generative AI developer assistant in production to all target developer teams in 2025. The company is also exploring additional use cases for generative AI, including additional internal and client-facing tools.

“Our focus is on emerging technologies and next-generation business capabilities,” says Powell. “We will work with product and business teams on how to enhance processes, improve metrics and efficiencies, and boost overall performance.”