How Progress governs AI-generated code at scale with SIG
Results at a glance
Security findings:
Flagged instantly
Time to validate new code:
3-15 Seconds
Potential yearly savings:
$ 2.25 Million per system
The need for secure and high-quality AI-assisted software development
Progress Software empowers over 100,000 organizations worldwide to achieve transformational success amid rapid technological change. Its platform enables teams to develop, deploy, and manage responsible AI-powered applications and digital experiences with agility and ease. Progress is a trusted provider of the technology, expertise, and vision enterprises rely on to succeed.
As artificial intelligence began reshaping how software is built, Progress saw both opportunity and risk. AI could dramatically accelerate innovation, but if left unchecked, it could also compromise reliability, security, and brand trust.
To capture the benefits while controlling the risk, Progress turned to Software Improvement Group (SIG), it’s trusted partner in continuous software portfolio governance.
The strategic challenges
1. Productivity of AI code at scale
Generative AI promised to dramatically accelerate development—but only if quality could keep up. Studies show that AI coding copilots can boost developer output by 26% on average, yet engineers often lose up to 19% of their time reworking AI-generated code.
2. Cost of poor code quality
Every hour spent fixing flawed or inefficient code carries a hidden cost. To quantify this cost, SIG uses a star rating system. This system provides independent, ISO 25010–aligned benchmarking, offering a clear 1–5 star quality rating.
SIG’s data shows that a 3-star rated system needs 0.7 extra FTEs per year, while a 2-star system requires 14, which can lead to over $2.25 million in avoidable spending. Progress wanted to reduce rework and protect margins as it expanded AI-driven delivery.
3. Security and trust in AI-generated code
As AI tools began writing and refactoring code, the risk of introducing unnoticed vulnerabilities grew by 40%. For a platform embedded in thousands of mission-critical applications, Progress needed real-time visibility into software security to ensure compliance, safeguard customer data, and maintain brand trust.
The solution
To tackle these challenges, Progress and SIG connected Progress’s development environment with SIG’s software governance platform, Sigrid®. This created a real-time control layer that checks and improves AI-generated code as it’s written. Instead of treating AI as an experiment, this approach made it a disciplined, measurable process: the same level of oversight that executives expect in finance or compliance is now applied to AI-assisted development.
"With quality and security checks built right into the workflow, developers can generate code, review it, and fix issues instantly, without ever leaving the IDE.”
The results
The joint initiative between Progress and SIG demonstrated that AI-driven development can meet enterprise expectations for security, speed, quality, and cost efficiency without trading one for the other.
Within weeks, the pilot delivered measurable business outcomes across all key dimensions of software performance:
| Metric | Before | After | Business Impact |
|---|---|---|---|
|
Security vulnerabilities |
Found post-testing |
Flagged instantly |
Avoids costly late-stage fixes; average breach now costs $4.88 M |
|
Time to validate new code |
Hours |
3-15 seconds |
Keeps teams in flow state, reducing context-switch loss |
|
Code quality issues detected |
5 per module |
None |
Eliminates debt at the source, preserving roadmap velocity |
|
Additional staff required for low-quality systems |
+14 engineers |
0 |
Potential $2.25M yearly saving per system |
Source: Progress whitepaper about SIG-joint Proof of Concept: ‘8 Steps to Unlocking Quality AI Code at Speed’
The takeaway for business leaders
AI is no longer an experiment, it’s becoming the backbone of how modern organizations build and deliver digital products. But as Progress and SIG proved, speed without control creates risk, and control without speed limits innovation. The answer lies in integrating both.
AI code governance is what makes AI sustainable.
The companies that pair automation with accountability will be the ones that innovate faster, operate leaner, and earn greater trust from their customers and regulators alike.
“AI is a powerful tool, but it needs oversight. With the Sigrid MCP server, developers can trust what the AI coding assistant produces, and ship better code, faster.”
What's next
SIG continues to help enterprises apply this model—combining speed, transparency, and control—so that AI becomes not just a productivity tool, but a measurable competitive advantage.
Learn how SIG helps enterprises like Progress go faster, responsibly.