AI in Platform Engineering: Assessing the Evolution
- Divine Odazie
- Jun 5
- 5 min read
The future of platform engineering isn’t just cloud-native; it’s now AI-native. Would you agree with that?
Whether your answer is yes or no, you surely would still agree that AI (Artificial Intelligence) is changing the way we build software.
According to Red Hat’s “State of Platform Engineering in the Age of AI” report, over 76% of organisations already utilise AI for various purposes, including documentation, code generation, and intelligent code suggestions.
It's not just noise; it's really happening.
If you’re driving platform engineering initiatives within your organisation, chances are you’ve asked yourself, ‘How would we take advantage of this innovation?’ If yes, then you’re not alone.
While the buzz around AI is loud, let’s take a step back and assess how far we have come in the world of platform engineering.
Current state of the ecosystem: Where AI meets platforms
The rise in popularity of Platform engineering today is a response to the complexity of the Software Development Life Cycle (SDLC). With Kubernetes clusters, CI/CD pipelines, observability stacks, and multi-cloud infrastructures, the average engineer is juggling more than ever.
AI, particularly powerful large language models (LLMs) and specialised AI assistants, has rapidly entered the software development sphere. Initially, adoption often occurred organically, with individual developers experimenting with publicly available tools like ChatGPT for various tasks or integrating coding assistants such as GitHub Copilot within their IDEs.
These early interactions, primarily focused on boosting individual productivity, have quickly gone beyond individual use cases, driving broader operational efficiencies and innovation.
The current state represents a meeting point. Platform engineering has matured, establishing Internal Developer Platform (IDPs) and standardised workflows as effective mechanisms for managing complexity and improving DevEx (Developer Experience).
Simultaneously, AI is offering a potent set of capabilities for intelligent automation, analysis, and assistance.
How AI is changing the game: Applications in platform engineering
As mentioned earlier, AI is offering practical solutions that can revolutionise platform engineering workflows. From platform development to infrastructure management and operations, there are several applications, including:
Code generation and optimisation.
Predictive analytics for platform health.
Automated security and compliance.
Cost optimisation.
Documentation and knowledge management.
All these applications come together to help platform teams achieve two overarching goals: elevating developer experience and gaining more buying power from executives.
Elevating developer experience (DevEx) through intelligent assistance
Ultimately, many AI applications in platform engineering converge on the primary goal: enhancing the developer experience. By automating tedious and repetitive tasks across coding, testing, and infrastructure management, AI directly reduces platform engineering load, freeing up valuable time and mental energy for more creative and complex problem-solving.
For example, AI assistants like the RunWhen Engineering Assistant can integrate directly into an IDP and can answer frequently asked questions, guide developers through standard platform workflows (e.g., requesting access, deploying a service), and offer initial troubleshooting suggestions based on error messages or observed system behaviour. This provides instant assistance and reduces the burden on platform engineers.
Unlocking platform insights for executives
Yes, developer experience is improving, but how will you communicate that to C-level executives?
We all know the lobbying needed to continually get buy-in from the top to invest more into improving a platform.
The data generated by an IDP and the underlying infrastructure is a rich source of operational intelligence, and AI provides the means to unlock it. AI tools can automatically ingest, process, and analyse data related to key engineering and platform performance indicators. With that, executives no longer need the platform team to generate up-to-date and plain-language reports on metrics to measure the IDP’s success.
LLMs could, in essence, act as a translator between the platform team and executives – stakeholder groups who often speak very different “languages” – making communication less frustrating and more efficient.
A double-edged sword? Challenges of AI in platform engineering
So far, you’ve seen how beneficial AI can be for a developer platform. But despite the attraction of these benefits, the path to AI integration is fraught with challenges that demand careful consideration.
Beyond hallucination – generating faulty configurations, misleading analysis, and bias – security is a major concern with AI. If we still face security threats with “non-AI” engineering, what about with AI? You remember when Wiz Research uncovered DeepSeek’s exposed sensitive data? The attack surface is significantly larger. We're not just dealing with traditional vulnerabilities, but also with new categories of threats unique to AI systems.
Practical implementation challenges also exist. Integration complexity arises when trying to seamlessly embed AI tools into existing, often intricate IDPs and development workflows. Tool incompatibility and the need for specialised integration efforts can be significant barriers. Compounding this is the skills gap; many platform teams lack personnel with the necessary expertise in AI/ML operations (MLOps), prompt engineering, and AI-specific security practices.
Important considerations for incorporating AI in platform engineering
Now you know how AI is changing the game for platform engineering and its challenges, how can you incorporate it into your existing developer platform? Or you’re building a new platform, how can you make it AI-native?
After identifying specific, well-defined use cases for AI integration, given the challenges of AI today, it is necessary to start with focused experimentation and maintain rigorous governance.
Open source platform engineering frameworks like Kratix offer the flexibility needed to conduct focused experiments. "Ready-made" platforms may have limitations in this area.
Kratix, as a framework, enables organisations to build composable IDPs through the definition and creation of “Promises (effectively platform component APIs) ”.
A Promise encapsulates all the required steps for deploying a service. This means you can encapsulate policies, permissions, and other guardrails for your AI experiments in a modular and reusable way, creating a “Golden Path” for your platform users.
Secondly, you should prioritise observability and maintain human oversight. The dynamic and sometimes opaque nature of AI models necessitates Human-in-the-Loop (HITL).
Platform teams need to establish clear metrics to track the performance and reliability of AI-powered features. This includes monitoring model accuracy, latency, and resource consumption.
A human-in-the-loop approach allows for learning and refinement of both the AI models and the integration strategies.
Now, how do platform engineers fit into all this?
What is the role of platform engineers in AI?
It's clear now that AI won’t replace engineers, but enable them to achieve more in less time.
Platform engineers who proactively develop skills related to AI integration, security, MLOps principles, and prompt engineering will be exceptionally valuable.
They will transition from primarily operational roles to becoming key strategic figures driving efficiency, innovation, and competitive advantage through the intelligent application of AI within the platform engineering discipline.
The platform teams of the future will likely require a blend of traditional infrastructure and DevOps skills combined with these new AI-centric capabilities.
Navigating the shift with confidence
AI is not a distant frontier. It’s the new baseline.
For teams already embracing “platform as a product”, the evolution is not about disruption but an augmentation.
At Syntasso, we believe IDPs must evolve to become intelligent platforms. Tools like Kratix provide the foundation for safely experimenting with AI integrations in a modular and policy-aware way.
Comentários