How Thredd uses AI to build software

With our journey toward building a next-generation issuer processing platform we have embraced Artificial Intelligence (AI) as a key enabler to accelerate development, enhance quality, and future-proof operations. Here is a summary of how we are using AI at various stages of our software development lifecycle:

Enhancing Development Velocity and Code Quality

By using AI, we transform our approach to software development through intelligent automation and augmentation:

• Code Generation and Review: We use tools like Amazon Q to automate code reviews, generate test cases, and produce technical documentation, ensuring consistency and quality in our Python codebase.
• Static and Dynamic Code Analysis: Security tools like Snyk and Sonar AI CodeFix help find vulnerabilities and suggest automated fixes, significantly reducing technical debt while keeping compliance with security standards.
• Legacy Code Modernization: While we initially explored using AI-powered tools (e.g., AWS Bedrock) for converting C# to Python, manual efforts proved more efficient. However, such tools stay part of our roadmap for future use cases involving legacy-to-modern technology transitions.

AI-Driven Operations and Support

We also plan to use AI to support our operational excellence by improving incident management and data handling:

• Incident Triage and Escalation: Large Language Models (LLMs) like those available in AWS Bedrock act as "switchboards," aiding in incident triage and escalation by guiding first-line support teams to resolution paths or right escalation points.
• Text-to-SQL Solutions: Leveraging AI for natural language-to-SQL query translation accelerates data insights, enabling non-technical teams to interact directly with databases via an AI agent interface.

Supporting Security and Compliance

We have been Integrating AI-based security solutions to ensure robust protection throughout the development and deployment process:

• Package and Library Scanning: Snyk finds vulnerabilities in Docker images and code libraries, automatically raising remediation pull requests.
• Code-Level Vulnerability Management: Sonar AI CodeFix provides proactive code fixes for found vulnerabilities, streamlining secure coding practices.

Reducing Development Costs with AI

AI potentially helps us achieve cost savings by automating tasks like code generation, reviews, testing, and documentation.

For example, developing a complex module manually might take 9.5 person-days. With AI tools like Amazon Q, this effort can be reduced to 4.5 person-days — a time reduction of over 50%.

These savings, when multiplied across the platform, allow us to distribute resources more strategically and accelerate our time-to-market.

The strategic integration of AI in our software development process positions our platform at the forefront of technological innovation, aligning with our broader business goals to modernize infrastructure, support growth, and deliver exceptional value to our customers.