Harness has just launched AIDA (AI Development Assistant), a generative AI assistant designed to streamline software development life cycle (SDLC) workflows.
Unlike traditional AI apps that focus mainly on code dev, AIDA looks at the whole SDLC – from code error resolution to security vulnerabilities and cloud cost governance.
Harish Doddala, VP of product management at Harness, said to VentureBeat:
“Our approach makes sure developers have AI-powered help throughout the SDLC, which we think is necessary to get the full benefits of AI in software delivery.”
Harness claims their generative AI tool can boost software engineering teams’ productivity by 30-50%.
It also offers automated detection and explanation of security vulnerabilities, drawing on its extensive training with data like common vulnerabilities and exposures (CVEs) and common weakness enumerations (CWEs).
Harness’s AI can generate explanations for security vulnerabilities and suggest remedies, saving time and effort on remediation.
This feature will help developers boost app security and keep code integrity during the SDLC.
Plus, any additional customization to meet code requirements will align with the company’s privacy and security policies. The AI solution is integrated with all Harness platform workflows, like CI, CD, cloud cost management, and feature flags.
Streamlining software dev with generative AI – it’s the way to go!
Doddala said AIDA gives developers quick pinpointing and insights into root causes to help them troubleshoot and resolve issues without manual log analysis. It analyzes log files, correlates error messages with known issues, and suggests fixes.
Plus, it uses AI to automatically identify security vulnerabilities and generate code fixes.
“What makes our solution special is its extensive training on known vulnerabilities and weaknesses,” said Doddala.
“This means AIDA can give developers actionable recommendations tailored to their codebase, boosting the software’s security.”
The company claims their AI tool helps developers manage cloud assets using natural language, so they can set up policies without having to manually code.
Doddala said they use a hybrid approach for data privacy and security, and are looking into using domain-specific data to train their models.
“We don’t send customer data without permission and safety protocols and encryption standards are in place. As for the LLMs, we’re looking at using pre-trained data with permissive licenses. We combine cloud APIs and our own LLMs,” he added.
What’s up next for Harness?
Doddala said AIDA’s initial release is just the start of its capabilities. In the coming months, it’ll introduce automated code reviews, AI-assisted CI/CD pipelines and AI-supported chaos engineering experiments.
Harness’s long-term vision for AI involves more generative AI and integrating it into software delivery.“
As generative AI evolves,” said Doddala, “it’ll continue reshaping software development, making it faster, more efficient and higher quality. We think AIDA will play a big role in shaping the future of AI-driven software development and giving developers these transformational capabilities.”