Clik here to view.
AI is rapidly transforming software development, but much of its value only focuses on increasing individual developer productivity. Code assistance brings welcome efficiencies to starting new projects, building out scaffolding and getting beyond the “blank screen” problem.
However, the true game-changer for software development is using agentic AI to realize the potential of platform engineering, enabling organizations to gain the maximum returns from their investments. This approach can dramatically improve code quality, reduce costs and increase developer productivity.
Agentic AI for Platform Engineering
More teams have attempted to adopt a platform engineering approach with tightly integrated tools and processes, but have struggled to achieve the full potential of these investments. Some of the biggest roadblocks are automating manual processes, scaling standardization efforts across teams, maintaining platform components and navigating complex and nuanced engineering environments.
Agentic AI is ideally suited for platform engineering. Many use cases — such as failure remediation, code review, test generation and coverage, documentation, security policies, network policies and change management — go beyond the scope of our mental models.
Code development in enterprise software is highly nuanced and contextual. Languages can have vastly different performance challenges. Junior developers may not have enough context to write prompts effectively. Security and compliance policies may also create unknown restrictions. No single platform engineer can fully grasp every security, network and application-layer concern across all these use cases.
Unlike traditional AI assistants that respond only to direct prompts, agentic AI has full context into a team’s software development infrastructure and can initiate actions based on triggers and states, making it the perfect complement to platform engineering frameworks.
Considerations When Implementing Agentic AI
Some of the key considerations leaders should think about when incorporating agentic AI into platform engineering workflows include:
Interoperability, Scalability and Reliability
- How will agents communicate with other agents, including across vendor product domains?
- How will agent “meshes” scale elastically and offer the same degree of predictability as microservices in their commission and decommission without performance impact?
- How will AI agents self-correct when they experience unexpected or unwanted results?
- How will they handle concurrency, multithreading, eventual consistency and other reliability and system failure arenas?
Security, Governance and Observability
- How will AI agents interact with existing network policies to define what they can and cannot access?
- How will agents interact with a multitude of data sources?
- How will the ingress and egress of agentic AI data work against existing data governance, security and privacy policies?
- How will AI agents’ telemetry data get collected, how will their performance be measured and what will remediation look like when they behave incorrectly?
Developer Workflows
- How will developers adapt their workflows to integrate with agent-based systems?
Enabling Velocity with AI Agents and Platform Engineering
One limitation teams face when using existing AI tools is the focus on individual productivity rather than team velocity. As AI agents mature, organizations can use these tools to infer and apply contexts across teams. These intelligent and adaptable AI agents go beyond fixed interfaces and preset workflows.
One area where I see rapid uptake for agentic AI is in the “tech mandatory” budget areas that most teams are committed to today, such as reducing technical debt, fixing security vulnerabilities, refactoring automation or infrastructure, and replatforming legacy apps. What all of these have in common is that they are rife with dense contexts and pose barriers to automation that agentic AI can remove.
For example, teams often create templates to standardize and automate processes at the platform level, such as a CI pipeline. This traditionally involves significant manual work to identify the right processes to target — those that are widely used, have repeatable steps and will have the most significant impact across teams. Agentic AI curtails those manual steps.
Rather than relying on human effort to identify processes for standardization, an agentic system can identify all Java-based projects from the past year, analyze the build processes across each and identify the best candidates for AI-based automation. The system can then create draft templates that the team can customize and build on.
The next stage of maturation in agentic AI systems is developing agentic mesh, a sophisticated ecosystem where AI agents can discover each other, collaborate and tackle complex challenges in previously impossible ways. These agents can monitor CI jobs, suggest process optimizations and implement those recommendations. They can also identify opportunities for cost optimization and directly adjust cloud resources to match demand patterns.
Platform engineering has delivered significant value, but many organizations have hit a ceiling in realizing its full potential. Agentic AI is the critical missing component that can uplevel platform engineering efforts by automating complex processes, applying contextual understanding at scale and enabling true team velocity rather than just individual productivity.
The post Agentic AI: The Missing Piece in Platform Engineering appeared first on The New Stack.
With full context, agentic AI can initiate actions based on triggers and states, making it the perfect complement to platform engineering.