How Dun & Bradstreet’s Google Gemini Code Assist is revolutionizing software development
Being at the forefront of innovation is important to Dun & Bradstreet, a top global supplier of business data and analytics. This dedication has included the company’s intentional strategy to enhancing its software development lifecycle through the integration of AI solutions for the past two years.
Although the company’s top priorities were security and development velocity, Dun & Bradstreet also faced numerous operational and productivity issues that are typical of many multinational corporations, such as:
- Onboarding new team members takes a long time.
- Separated expertise in legacy codebases
- Low coverage of tests
- The challenges of modernizing applications
Dun & Bradstreet realized it needed to address the usual development lifecycle problems that most enterprise engineering teams encounter by adopting a comprehensive “people, process, and tools” strategy to speed up software development. They looked to AI support as the foundation for this new endeavor.
The software development lifecycle’s future partner search
Since Dun & Bradstreet provides information that has the power to influence markets and economies, their standards for technology tools were as high as those of the government officials and financial professionals they work with.
Taking into account both open-source and commercial choices, Dun & Bradstreet carried out a comprehensive review process to determine the optimal partner and coding aid tool. Because of the Gemini model’s strong security features, smooth interaction with their current development environment, and performance, the organization finally decided on Google Gemini Code Assist.
The Google Cloud team worked together with the development teams at Dun & Bradstreet to deploy Gemini Code Assist. The team’s developers actively participated in the tool’s design and customization to make sure it suited their unique requirements and work processes.
Google’s security posture was one of Dun & Bradstreet’s main areas of focus. It was necessary to have both guardrails to guarantee the security of machine-generated code and the highest level of privacy protection while integrating AI into the development process. The following elements of Google’s security knowledge and direction gave Dun & Bradstreet the confidence to proceed:
- Google is able to thoroughly check and filter all source code samples used in model training because Gemini models are developed internally.
- Trust and verify: Developers can oversee the model outputs with human oversight by integrating it into an organization’s current coding and review lifecycles, eliminating the need to learn a completely new system.
- Google and Snyk’s collaboration offers more automatic security scanning choices that cover both human-written and AI-generated code.
- The architecture and design choices for Google Gemini Code Assist are based on Google’s AI Principles. Fine-grained administrative controls to stop unintentional data leaks, encrypted logging, and single-tenant storage of customer code references are examples of privacy and security measures.
- Google’s policies regarding indemnity.
A change in action
The decision to proceed with Google Gemini Code Assist was made by Dun & Bradstreet in October 2024. The solution is now beginning to be implemented by additional teams around the company. Code Assist’s user-friendly UI and thorough documentation have made adoption easy.
Implementing an incubation program can help large organizations overcome possible adoption and technological obstacles. For example, the Dun & Bradstreet team determined that teams needed to be made aware that coding assistants are there to support developers as partners, not as substitutes.
Now that the rollout has started, Dun & Bradstreet is discussing the reasons behind their decision to use Google Gemini Code Assist.
Enhanced developer efficiency: Gemini Code Assist‘s AI-driven code completions and recommendations have drastically cut down on the amount of time developers spend creating code. By automating tedious activities, the program has allowed engineers to focus on more intricate and imaginative areas of their work.
Better code quality: Early in the development process, Dun & Bradstreet’s engineers were able to identify mistakes and possible problems with Google Gemini Code Assist‘s automated code review and linting features. As a result, there are now a lot fewer defects and better code overall.
Simpler modernization of applications: Transforming Spring apps to Kotlin saved a substantial amount of time.
Enhanced developer productivity: Initial internal data indicates a 30% rise in developer output.
Onboarding of developers: Gemini Code Assist’s real-time advice and assistance have allowed Dun & Bradstreet’s new developers to ramp up more quickly.
Enhanced information sharing: The development teams at Dun & Bradstreet now have a culture of knowledge sharing because to Google Gemini Code Assist. The program makes it simpler for developers to work together and learn from one another by offering code examples and best practices.
Taking the lead in AI
With its sophisticated AI capabilities and intelligent code assistance, Google Gemini Code Assist has been a useful tool for Dun & Bradstreet’s engineers.
The business has been successful in increasing velocity, enhancing software quality, and preserving its market advantage. Software developer lifecycles are significantly improved by Google Cloud and Gemini, according to businesses like Dun & Bradstreet.
