R Systems International Limited, a global systems integrator, this week revealed it is building a coding practice based on Cursor, an artificial intelligence (AI) coding tool developed by Anysphere.
The systems integrator plans to train more than 1,000 engineers on how to use Cursor to both build new applications and modernize existing legacy systems.
R Systems CTO Srikara Rao said Cursor was selected over other AI coding tools because of its ability to not just generate code but also understand the context of the environment where that code needs to run. In contrast, other AI coding tools will generate code that doesn’t actually run in a production environment because the tool itself lacks any context about the underlying codebase, he added.
As part of the alliance, R Systems will establish a dedicated Co-Innovation Lab focused on advancing AI-first software engineering practices, including a reusable prompt repository and a knowledge base of best practices. The lab will also define a repeatable software development lifecycle framework, including playbooks, that injects AI into coding, testing, documentation, project management and DevOps workflows. Additionally, Cursor will be incorporated into OptimaAI, a proprietary AI platform that R Systems previously rolled out.
The overall goal is to improve delivery velocity by ~30%, reduce defect density by 25%, and halve engineer onboarding time.
Cursor is a code editor based on a fork of VS Code project. It makes use of large language models (LLMs) such as Claude from Anthropic and GPT from OpenAI to write code, fix bugs and even explain codebases in plain English. Organizations that have adopted Cursor include Rippling, Adobe, and Nvidia to streamline software engineering tasks and workflows.
Cursor can also autocomplete entire functions, refactor code, suggest performance improvements and respond to prompts right inside the editor. It also indexes the entire code base and then relies on embeddings to retrieve the right files.
There is also an Agent Mode that can plan and carry out a sequence of steps to complete complex tasks, which plays a crucial role in reducing, for example, the number of consultants that would otherwise be required to modernize a 20-year-old project, said Rao. While some of that savings may drop to the bottom line, more of it will be allocated to funding additional application development projects, he added.
It’s not clear to what extent DevOps teams have embraced AI tools, but there is clearly a battle underway between a raft of newcomers such as Anysphere and incumbent providers of legacy tools and platforms. While many organizations are experimenting with a wide range of AI tools, the degree to which they might be willing to replace incumbents that are also working on AI initiatives will vary widely.
Regardless of approach, the one thing that is certain is that application development is being transformed utterly in the age of AI. The next major challenge will then be determining how best to expand existing DevOps pipelines to accommodate the volume of code that is now being generated using any number of AI coding tools that are readily accessible to any developer who cares to experiment with them.
