State of the Art of Platform Engineering • Abby Bangser & Charles Humble • GOTO 2026

2026-06-22 · Source: GOTO Conferences · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, extended

Summary

The "State of the Art of Platform Engineering" discussion with Abby Bangser and Charles Humble highlights the discipline's evolution towards self-service internal developer platforms. While understanding the value of centralization for cost savings and compliance is widespread, many organizations struggle with execution, often rehashing centralized operations. An emerging architectural pattern involves internal marketplaces to scale platform building, as observed at KubeCon 2025 and 2026. Key aspects include treating the platform as a product with a clear scope, avoiding "platform facades" created by portals and CI/CD pipelines that lack true self-service and longevity, and recognizing signs of platform decay. The conversation emphasizes making a business case for maintenance by tracking cost and opportunity impacts. Critically, AI is transforming platform engineering by making AI agents first-class consumers, driving the need for API-first, deterministic capabilities and robust guardrails to enable confident AI adoption. The future of platform engineering hinges on achieving compliance and coherency at the scale of AI consumption.

Key takeaway

For AI Architects and Platform Engineers evolving internal developer platforms, prioritize delivering API-first, deterministic capabilities with robust guardrails. This is crucial for enabling trusted, scalable AI agent consumption and ensuring compliance, shifting the platform's core value. Justify platform investments by tracking cost savings and opportunity costs, rather than solely developer experience, to secure necessary funding. Failure to provide these foundational elements will constrain AI adoption and accelerate platform decay.

Key insights

Platform engineering must provide self-service, API-first, deterministic capabilities to enable scalable, compliant AI agent consumption.

Principles

• Centralization with self-service unblocks value streams and ensures compliance.
• Treat a platform as a product with clear scope, including feature deprecation.
• Platform capabilities should be implementation agnostic and fleet managed.

Method

Begin with high collaboration to understand user needs, then facilitate across multiple consumers, progressing to automated on-demand APIs.

In practice

• Identify platform decay by observing development teams "cutting corners" or avoiding friction.
• Design platform APIs to serve both human and AI agent consumers effectively.

Topics

• Platform Engineering
• Internal Developer Platforms
• AI Agents
• API-First Design
• Platform as a Product
• Technical Debt

Best for: AI Engineer, MLOps Engineer, AI Architect

Related on AIssential

• Platforms for Humans and Machines: Engineering for the Age of Agents — Juan Herreros Elorza — Platform engineering principles like self-service and API-first design are crucial for enabling AI agent productivity.
• Designing AI Platforms That Scale: A Practical Blueprint — Centralize AI governance and observability to enable federated development and deployment, balancing control with speed.
• Platform as a Product: Delivering Value While Balancing Competing Priorities — Abby Bangser's talk at GOTO Copenhagen advocates for treating software platforms as products, not just code, to deliver sustainable value by balancing engineering, design, usability, security, and internal customer needs.
• Podcast: Engineering Stable, Secure and Scalable Platforms: A Conversation with Matthew Liste — Platform engineering prioritizes stability, security, and scalability, using customer journeys to manage risk and adapt to AI's accelerating impact.
• The Organization Is the Bottleneck — Organizational maturity and existing engineering foundations dictate AI coding agent success, not the tools themselves.
• I don't have time to build these things, will you? — AI agents enable tackling ambitious, foundational software problems previously deemed too complex or expensive.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by GOTO Conferences.