How engineers can use AI to reduce coding time and improve code quality

The impact of AI on our engineering practices has been transformative. Our software engineers are using AI to improve code quality, achieve faster development cycles, and learn new skills.  

Take a look at early results from our AI Assisted Engineering program:   

• Faster Development Cycles -  reduced coding time by 30%  

• Improved Code Quality - 61% of our engineers experienced improvements in code quality, reducing defects and technical debt over time.
• Knowledge Transfer: ~33% of engineers reported learning best practices through AI suggestions, fostering professional growth and elevating team-wide standards.

Here’s how our engineers are using a governed set of AI tools to achieve those results: 

1. Handling essential coding tasks that are repetitive and time-intensive, so our engineers can focus on product innovation and strategic problem-solving. Routine activities like code generation, bug detection and debugging, maintaining legacy codebases, and optimization.  
2. Creating consistency in code quality, which is particularly important for large companies with engineers at varying skill levels. Code consistency decreases the risk of technical debt, lower maintainability, and frequent regressions. Tools like GitHub Copilot provide syntactically correct, contextually accurate code suggestions that speed up development and serve as guardrails that help engineers produce cleaner, more maintainable code.
3. Exposing newer engineers to best practices. AI is bridging knowledge gaps by giving junior engineers, and those who are new to specific domains, exposure to industry best practices. Learning as a continuous loop! 
4. Accelerating innovation by helping with experimentation and iteration tasks - like prototyping, performance optimization, error detection - that are typically constrained by traditional development timelines.  
5. Scaling our global engineering team, by removing knowledge silos and uneven performance across engineering functions. AI tools are providing a scalable way to standardize workflows and ensure consistent performance, whether the team comprises newly onboarded IC1 engineers or experienced architects.

Adoption is key to the success of any transformative program. Murali Swaminathan and our Freshworks engineering team followed an ‘Embrace, Enable, Expand’ framework to ensure a smooth integration of AI tools into our engineering processes. Here are steps you can take to start using AI in your own coding environments. 

1. Embrace: Build Awareness and Confidence 

To introduce new AI tools and cultivate a mindset of openness and curiosity among engineers, take the following steps:  

• Showcase Potential: Conduct targeted workshops and live demonstrations to showcase the tangible benefits of AI. These sessions should highlight real-world scenarios where AI could save time, improve code quality, and assist engineers in learning best practices.
• Address Concerns: Some engineers are skeptical about relying on AI tools, worrying about accuracy, relevance, or even job displacement. Hold open forums and Q&A sessions to address concerns and further show how AI serves as an augmentation tool—not a replacement.
• Pilot Teams: Identify early adopters and onboard them as pilot teams. These teams can provide invaluable feedback and serve as internal champions, demonstrating success stories that inspire others.

2. Enable: Equip Engineers for Success

Provide engineers with the resources, training, and support needed to effectively use AI tools in their workflows. This can come in multiple formats: 

• Tailored Training: Recognizing the diversity in engineering roles and expertise levels, design role-specific training programs. For example:
• Hands-On Tutorials: Familiarize engineers with the tools. Host real-world coding scenarios and guided sessions to build trust in AI-generated outputs.
• Documentation and FAQs: Create a comprehensive knowledge base, including best practices for AI tool usage, troubleshooting guides, and success case studies.
• Feedback Loops: Encourage engineers to share their experiences through structured feedback forms - this will help to iteratively refine the tools and training programs.

3. Expand: Scale Adoption Organization-Wide

Drive broad adoption of AI tools across all teams and functions, to ensure sustained impact at scale. Here’s how you can achieve this: 

• Phased Rollout: After successful pilots, expand the program incrementally, ensuring that each team received adequate support and training during onboarding.
• Cross-Functional Integration: Tailor AI tools and integrate across various engineering functions.  
• Governed Usage: Establish governance policies to ensure responsible use of AI tools, focusing on data privacy, compliance, and ethical considerations.
• Support Ecosystem: Check in with  internal AI champions and helpdesks regularly to ensure teams have access to continuous support.

This approach worked for us and our engineers quickly saw tangible benefits. The feedback loops enabled iterative improvements, making the tools increasingly effective as the program scaled. These AI tools are now an integral part of our engineering ecosystem, driving both short-term gains and long-term transformation.