
What if the principles that transformed software development over the last decade could be the key to successfully implementing AI in your organization?
In the latest Mik + One podcast, Dr. Mik Kersten, Chief Technology Officer at Planview, sat down with Patrick Debois, known as the “godfather of DevOps,” to explore this question.
Their conversation reveals how DevOps practices—particularly experience with managing unpredictable systems, scaling infrastructure, and evolving testing approaches—are proving crucial for AI implementation.
From organizational structure to developer roles, Debois shares insights on why DevOps experience gives teams a unique advantage in navigating the challenges of enterprise AI adoption.
Who is Patrick Debois?
Patrick Debois is credited with coining the term “DevOps” and has been instrumental in shaping how organizations approach software development and operations.
With a career spanning various roles from developer to ops person to manager, Patrick has consistently focused on building bridges between different organizational functions. Today, he’s applying these same integration and automation principles to the challenges of implementing generative AI in enterprise environments.
Podcast Key Takeaways
- DevOps principles of resilience engineering and observability are essential for AI system success.
- Platform teams should centralize AI infrastructure including model access, tracing, and governance.
- Organizations need new testing frameworks that incorporate AI-based quality assessment and gradual releases.
- The future of development involves balancing automation with maintaining system understanding.
Best Clip
Patrick discusses the role of automation and developers, arguing that automation is beneficial but comes with pitfalls like over-reliance and complacency, which can make certain tasks harder.
3 Ways DevOps Addresses AI Challenges
“To get the first thing out of the door, you need a use case that brings the value. If you’re not iterating on that—if you’re immediately thinking about all the problem space and on the solution—that’s a problem in itself.”
Patrick Debois
Transitioning to AI-enabled development presents significant challenges, but DevOps practitioners are uniquely positioned to address them. Here’s why:
Managing unpredictability and change
The rapidly evolving AI landscape, with its shifting tools and frameworks, creates significant uncertainty for organizations. DevOps teams excel here because they’ve mastered handling unpredictable systems and building resilient architectures.
Patrick emphasizes how DevOps practices of continuous integration and flexible deployment enable organizations to easily switch between different AI providers and frameworks rather than becoming locked into specific solutions.
Läs nästa: Att bygga en AI-strategi för att lyckas med digital transformation
Building effective feedback loops
AI systems introduce new complexities in testing and quality assurance that traditional approaches can’t fully address. DevOps teams are already skilled at what Patrick calls “dealing better with failure”—creating robust feedback loops and monitoring systems.
This experience translates directly to AI implementation, where production feedback becomes even more critical than in traditional software development.
Läs nästa: Mastering Integration Automation: How to Build, Test, and Deploy with Continuous Integration
Scaling infrastructure and operations
Organizations struggle with decisions about building versus buying AI infrastructure and managing data accessibility. DevOps teams bring valuable experience in scaling infrastructure and managing API-driven systems.
Patrick highlights how DevOps practices around platform teams and shared services can help organizations build centralized AI infrastructure while maintaining team autonomy—a critical balance for successful AI implementation.
Läs nästa: Trends in AI: Should You View Data Like a Product?
Changing the Future of Development
We’re all moving from more producing to more reviewing. We’re moving from a diff[erence] to an ops person who deals with reviewing and understanding the implications of what’s being generated.
Patrick Debois
Patrick highlights that while AI will transform how we develop software, the fundamental principles of DevOps—especially those related to handling failure, ensuring resilience, and maintaining observability—will become even more critical.
Organizations that can successfully apply these principles while adapting to AI’s unique characteristics will be best positioned to leverage this technology effectively.
Want More?
Listen to the full episode to hear:
- How to build effective AI feedback systems that prevent costly mistakes while encouraging innovation
- Practical strategies for measuring AI implementation success beyond traditional metrics
- Patrick’s insights on avoiding common AI adoption pitfalls, drawn from his extensive DevOps transformation experience
Check out the Mik + One archives to hear conversations with other industry experts. Be sure to subscribe to learn when new episodes are released.
Episode 60: Robin Yeman and Dr. Suzette Johnson on Implementing Lean, Agile, and DevOps Principles
Episode 42: Manuel Pais on Organizing Business and Technology Teams for Fast Flow
Episode 44: Geordie Henderson on Implementing Organizational Design and Innovation to Scale High-Performing Teams