Why AI demands mature DevOps: The case for The DevOps Standard Official Book
Marc Hornbeek (DevOps_the_Gray), Senior DevOps Advisor, Ambassador for DEVOPS INSTITUTE
AI is transforming the speed and scale of software development, but many organizations have not adapted their delivery systems accordingly. For Marc Hornbeek, this growing imbalance inspired the creation of the new official DEVOPS INSTITUTE® book, The DevOps Standard: The Vendor-Neutral Model for the AI-Driven World! Hornbeek has dedicated years to helping organizations improve their DevOps maturity and believes the industry needs a more prescriptive, vendor-neutral approach. The new book introduces the DEVOPS INSTITUTE® Operating Model that combines the Nine Pillars of DevOps practices with a four-layer DevOps Blueprint. This model integrates DevOps, DevSecOps, SRE and Platform Engineering to help organizations create more balanced and sustainable value streams.
For a long time, I have been a proponent of a prescriptive DevOps approach. What prompted the need for a book now was the realization that the industry needs an official reference from a recognized body like DEVOPS INSTITUTE.
The biggest catalyst has been how organizations are adopting AI. While AI allows organizations to generate code very quickly, that is only one part of the software value stream. If integration, testing and security don't keep pace, the value stream becomes unbalanced.
I’ve seen companies struggle to achieve mature DevOps for years, with or without AI. There are many moving parts, and organizations can easily get ahead in one area without bringing the rest of the system along. AI makes those issues much more apparent because it increases the speed at which change is introduced.
Think about it like building a fast car. If you want a car to travel faster, you can’t just change one component and expect everything else to work. You have to design the whole vehicle for higher speeds. It’s the same with DevOps. You need an operating model that keeps the different elements in balance.
This is the book's primary purpose: to provide a prescriptive approach that helps organizations ensure their DevOps is mature enough to keep up with AI-generated code.
Why DevOps must work as a system
That balance is difficult to achieve if DevOps is treated simply as a culture initiative, a team structure, or a collection of tools. These are all important parts of DevOps, but none represents the whole system.
DevOps is an organization-wide system for delivering value. It brings together people, practices, platforms, security, reliability, governance, measurement, and continuous learning across the entire value stream.
It also needs to connect the capabilities now central to modern software delivery. DevOps, DevSecOps, SRE, and Platform Engineering should not be treated as separate disciplines. Each contributes something important, but their value comes from how they work together across the delivery system.
The book therefore defines the DEVOPS INSTITUTE Operating Model rather than focusing on one technology, team or methodology. The aim is to provide a structure that connects these elements to reliable business outcomes.
Building a balanced DevOps model
The DEVOPS INSTITUTE Operating Model combines the Nine Pillars of DevOps practices with a four-layer DevOps Blueprint: Elastic Infrastructure, CI/CD Pipeline, Application Release Orchestration, and Value Stream Management.
The Nine Pillars define the practices that need to mature, while the Blueprint shows where those practices operate across the delivery architecture. The model integrates DevOps, DevSecOps, SRE, and Platform Engineering capabilities. AI is infused throughout as part of the modern software delivery environment.
Together, these elements connect people, processes, technology, governance and measurement with reliable business outcomes. Organizations can use the model for education, implementation, assessment, governance and continuous improvement, while adapting it to their technology, scale, risk and regulatory context.
Where existing approaches fall short
Many DevOps frameworks, models, and bodies of practice are already available. Many published approaches are vendor-specific. This can be difficult for organizations that use a mix of vendors and technologies, as vendor-specific approaches often focus on optimizing one part of the delivery environment. In contrast, mature DevOps requires balance across the whole value stream.
Some organizations have invested heavily enough to develop mature DevOps capabilities, particularly the largest technology companies and major corporations. However, only a relatively small number of organizations have achieved what I would call mature DevOps.
Influential books have also shaped the industry. The Phoenix Project, The DevOps Handbook and Accelerate are extremely valuable, and my own Engineering DevOps book, published in 2019, addresses many of these issues. What they don't provide is a complete prescription for a DevOps operating model and a structured path to maturity. They describe values, principles and practices, but they do not provide that complete operating model.
This is the gap the book aims to address. Organizations shouldn’t have to experiment with how to achieve mature DevOps every time. They need a recognized structure to guide them toward mature DevOps.
Finding the constraints
The book gives organizations a starting point for understanding their current state. The assessment model helps determine where an organization is today and where its gaps are. From there, value stream mapping can help identify the biggest constraints affecting performance. This matters because organizations can otherwise end up trying to improve everything at once.
Once the major constraints have been identified, the book suggests practices that can help address those gaps. Organizations can then use a focused 90-day improvement effort to test those changes, measure the outcomes, capture reusable practices, and scale what has been proven. The book is not intended to be a massive implementation manual telling every organization exactly what technology to deploy or how to configure every component. Instead, it provides a framework and a standard, with enough implementation indicators to give organizations meaningful guidance without prescribing every implementation detail.
Why AI-generated code needs mature DevOps
AI-generated code makes the need for this maturity especially clear. The 2025 DORA State of AI-assisted Software Development report found that AI acts as an amplifier, magnifying an organization’s existing strengths and weaknesses. That makes the underlying software delivery system increasingly important as organizations adopt AI for development.
Testing, integration and security controls still have to work, regardless of who or what generated the code. When they don't, organizations can put AI-generated code into production without adequate validation. This can contribute to significant technical debt, and organizations may not even have enough people to carry out all the necessary technical reviews.
This is where good DevOps fills a major gap. AI doesn’t remove the need for DevOps; it makes mature DevOps more important. The faster an organization can generate software, the more important it is to have the practices and controls needed to validate and deliver it safely and reliably.
Building AI into the value stream
The DEVOPS INSTITUTE Operating Model treats AI as part of the broader delivery system. Teams need to manage AI-generated code, artefacts and decisions within the same governed delivery system as human work. Organizations therefore need appropriate practices for provenance, validation, security, traceability, risk-based controls, production observability, and human accountability.
Bringing these capabilities together matters even more as AI becomes more deeply embedded in the value stream. This means organizations need an operating model that clarifies what is being introduced, how it has been validated, what risks are involved, and who remains accountable.
What mature DevOps looks like
The book is aimed at executives, transformation leaders, DevOps and DevSecOps practitioners, SREs, platform engineers, developers, architects, assessors, educators, consultants and coaches.
What I hope readers take away is that they should manage DevOps as a coherent, measurable, secure, reliable, and AI-enabled operating system for value delivery. As AI changes the speed and scale of software development, organizations need the maturity to turn that increased velocity into reliable business value. The new official DEVOPS INSTITUTE book provides that starting point.