CodeScene is extremely proud to announce our AI Generated Code Refactoring tool and to invite you to our Beta testing program! Meet our AI Coding Assistant.
You'll get first access to our proprietary technology that is the only generative AI coding assistant with up to 97% accuracy.
Not only did we train a generative AI to write code for refactoring, we created a secondary AI to fact check it.
AI Coding Revolution? Yes. But AI is nowhere near replacing humans in a coding context.
AI coding revolution is here, and there are plenty of tools that might be ok for code writing inspiration - like CodeWhisperer, Copilot and many more. But the burden is still on developers to ensure that the AI generated code is correct, that it actually works.
And most importantly, to ensure that by adding AI generated code to the codebase the developer is not introducing new bugs or more problems down the road. Can you really trust AI coding assistants to do quality work for you? Or are you in the risk of degrading your code quality?
What is mainly time consuming for developers is working with the existing code, with the big bottleneck being understanding what the code does in the first place. By refactoring code, you shorten the cycle to understanding. Thus far, this isn’t an area addressed by generative AI, and current usage of AI assisted programming is primarily for writing new code.
That's why we've created our new AI automated refactoring tool.
Let's have a look at some important benefits:
- CodeScene already has the only code-level metric with proven business impact – the powerful Code Health Analysis, which we built our automated code review on. That way, you know what needs to be refactored as well as why.
- CodeScene is the only tool on the market which can identify and prioritize refactoring targets based on how you – as an organization – work with the code.
With our AI refactoring tool built on top of our core product, you can now automate refactoring of unhealthy code which is negatively impacting your overall code quality and contributing to technical debt.
Auto-mitigating technical debt and automated improvements to existing code
We know that paying down technical debt improves future performance and delivery efficiency.
You'll actually cut down on time and resources needed for refactoring and improving code quality and maintainability, essentially getting double the impact of other generative AI code writing tools. AI should take care of the hard parts so devs can focus on innovating.
Can we back up our innovation? Yes we can.
Recently we've published a benchmark study investigating AI support for improving existing code.
As the authors of this paper state: "The majority of a developer’s time isn't writing but understanding and maintaining existing code." That's where we have the real gain.Some of the findings by Adam Tornhill, Markus Borg and Enys Mones:
- They benchmarked the performance of the most popular Large-Language Models (LLM) on refactoring tasks for improving real-world code. Their findings show that existing AI solutions only deliver functionally correct refactorings in 37% of the cases.
- As a response to the poor performance of LLMs, they then introduce a novel innovation for fact-checking the AI output and augmenting the proposed refactorings with a confidence level. By rejecting incorrect solutions, 98% of the remaining AI-generated refactorings improve the code while retaining the original behavior.
Join the beta waitlist and get early access to ACE
