Software Improvement Group launches Sigrid® Axis to control what AI coding agents produce
Deterministic checks against the world’s largest software quality benchmark, rooted in international standards, so agents create business impact instead of rework.
Amsterdam, 2 October 2026–Software Improvement Group (SIG) today launched Sigrid Axis. This platform works with AI coding agents to produce secure, maintainable code that respects the architecture it lives in. Each check runs against the architecture of the individual system, repository, or component the agent is working in, not just an organization-wide set of rules.
SIG’s existing software portfolio governance platform is now called Sigrid Core. Both share the same backbone: Core continuously governs the entire software portfolio, sets the standards, and enforces them in the CI pipeline; Axis holds agents to those standards while they work.
Organizations are moving from AI-assisted coding toward autonomous agents that write code. Their controls are typically either generic or tailored to a single system and often lack an external quality standard to check against.
“Many teams rely on careful prompting or a second AI tool to check the first, but that cannot be the only safeguard. What they often get instead is rework and review cycles that eat the speed the agents delivered in the first place. Trust has to be built in as the code is written, against established standards, combined with what matters to the individual organization,” said Luc Brandts, CEO at Software Improvement Group.
Many tools will tell an organization what good looks like; Sigrid assesses code against deterministic rules, calibrated on more than 40,000 real systems and over 600 billion lines of code. Standards have been refined over 25 years in the world’s only ISO/IEC 17025-accredited software quality lab. On top of this, Sigrid presents findings and fixes in the context of the organization’s own priorities and architecture design.
Sigrid Core: Continuously govern the entire software portfolio and set high standards
Sigrid Axis does not replace what is now known as Sigrid Core.
Sigrid Core is the control center for an organization’s entire software portfolio from a security, productivity, compliance, ROI, and AI-impact perspective.
Sigrid Axis: Hold agents to those standards while they work
Sigrid Axis is built for agentic engineering. It anchors agents in the organization’s architecture and plans their work, then catches problems as code is written and improves existing code. That is what stops technical debt from compounding, and compounding debt slows down humans and agents alike in understanding the code and building on it.
Axis connects through MCP, the open standard the major AI coding tools speak, so it runs inside the tools teams already use, including Claude Code, GitHub Copilot, Cursor, and ChatGPT.
- Guardrails
Real-time checks on every change an AI coding agent makes, with violations fixed before they are committed. - Auto-fix
Find and fix what matters most in the existing codebase, logging every change for a complete audit trail. - Dashboard
Tracks AI adoption across the portfolio, and its effect on productivity, quality, and cost.
In recent beta testing with existing SIG customers, review comments per pull request fell from 3.9 to 0.9 in one case with Guardrails switched on. “I particularly value the relational graphs and grounding truth that Sigrid Axis provides for code-based intelligence,” another beta customer said.
Used together, Core and Axis are the most complete way for enterprise organizations to govern the large software portfolio they already run and adapt it for the agentic era. Axis can start with a single system, inside the coding tools a team already uses. Core is the bigger footprint, covering every system in the portfolio on the same standard, which is how organizations move from governing agentic development to governing the entire portfolio their business runs on.
“Adding reasoning to help an agent make better choices is not new. Grounding that agent in objective standards and checking every file it writes against the world’s largest software quality benchmark before a reviewer sees it, is. Trust is what decides how much you let an agent do on its own, and that is what we built Axis to give you,” said Jasper Geurts, Chief Technology Officer at Software Improvement Group.
SIG has always supported independent, fact-based decision-making. The quality standards behind that do not change with this launch, and neither does the enforcement: Sigrid Core already applies those standards in the pipeline as part of the definition of done. What changes is where they apply. The same standards now run inside the loop that produces the code, and Sigrid Axis fixes what it finds before it ships.
About Software Improvement Group
Software Improvement Group (SIG) empowers organizations to govern the software their business runs on. Through complete portfolio analysis, benchmarking, and strategic guidance, SIG helps companies build better software and adopt AI without losing grip on cost, speed, and risks.
Sigrid®, its software portfolio governance platform, is offered as Sigrid Core and Sigrid Axis. AI makes writing code faster than ever, and harder to trust. Sigrid Core is built on SIG’s benchmark data and quality models, showing organizations whether their software is sound and what it will cost them if it isn’t. Sigrid Axis builds on that same foundation as the deterministic layer for probabilistic AI and works with AI coding agents to produce code that is secure, maintainable, and true to the architectural design.
SIG operates the world’s largest software quality benchmark, consisting of over 600 billion lines of code across more than 40,000 systems and 300+ technologies. It is also the world’s only ISO/IEC 17025-accredited laboratory for software quality analysis, and co-developed standards such as ISO/IEC 5338, the global standard for AI lifecycle management, and prEN 18282, the forthcoming official AI Act security standard.
Founded in 2000 and headquartered in Amsterdam, SIG has offices in North America and Europe, and partnerships across the world for global support.
Combining renowned quality models with technology and over 25 years of industry-leading expertise, SIG is the global authority on continuous software portfolio governance.
For more information, please visit Software Improvement Group‘s website or social media channels.
