Google’s Gemini enterprise coding assistant shows enterprise-focused coding is growing

Google Cloud’s newest feature, Gemini Code Assist Enterprise aims to compete with GitHub’s enterprise-focused coding platform to explain local codebases and get more security. 

Gemini Code Assist Enterprise, formerly Duet AI, lets developers code faster because it understands their organization’s codebase, has a large context window, and allows for customization. Developers can access the assistant for $45 per month per user or $19 monthly with a yearly subscription.

“Developers can stay in flow state longer, bringing more insights directly to their IDEs, while also completing complex tasks like upgrading a Java version in an entire repo,” said Ryan J. Salva, senior director, Developer Tools and Operations, Google Cloud in a blog post. “This means developers get to focus on creative problem-solving, leading to greater job satisfaction while you get a faster time-to-market, gaining a competitive edge.”

The platform offers code suggestions based on local codebases. Google said the large context window helps developers “generate or transform code that’s more relevant to your application.”

The coding assistant can connect directly to other Google Cloud services like Firebase, Databases, BigQuery, Colab Enterprise, Apigee and Application Integration. Salva said this is to meet developers where they are since “the more services it touches, the faster your builders can create and deliver applications.”

The code customization is based on internal libraries so Code Assist can help make custom code suggestions. It will index GitHub and GitLab libraries and support self-hosted libraries early next year.

“A code assistant dramatically reduces the time to ramp on new technologies and incorporates the nuances of an organization’s coding standards into the suggestions it provides,” Salva wrote.

But Google’s biggest selling point for the coding assistant is its enterprise-grade security. It extends Google’s promise that it won’t use customer data to train its Gemini models. It also promises that users have complete control over which repositories the code assistant will index, and they can purge data anytime. Google will also offer indemnification — legal cover for any potential lawsuit — for any code generated by Gemini Code Assist Enterprise. 

Enterprise-focused coding assistants

Coding assistance, of course, is nothing new for generative AI. But as more enterprises hope to integrate coding assistants into their technology stack, providers hope to tailor their offerings to them. 

GitHub released an enterprise-focused Copilot called GitHub Copilot Enterprise in February, largely offering similar features. Oracle’s coding assistant focuses on Java and SQL enterprise applications. Other companies like Harness also came out with coding assistants that give real-time suggestions and target businesses. Harness’ assistant is built off Gemini. 

Google’s entering the fray underscores the increasing competition in coding assistants and the need to make enterprise-specific solutions even for a task most chatbots can readily do. Moving coding assistants from separate chatbots and integrating these into developer environments or in Google’s case,e other channels gives flexibility to companies looking to improve productivity. The more developers can quickly test code and maybe fix bugs on local codebases, the faster companies can move and deploy applications.