The annual Stack Overflow Developer Survey is always a treasure trove for understanding where the developer community is headed—and, more importantly, where the pain points lie. This year’s 2025 survey delivers a clear message for platform engineers: AI is here, adoption is high, but trust is slipping. That gap between usage and confidence is where platform engineering can — and must — step in.
AI Everywhere… But Not Trusted Everywhere
The big headline from the survey is that 84% of developers are using or plan to use AI tools in their workflow. For platform teams, that means the AI genie is already out of the bottle — developers are bringing these tools into the environment whether they’re officially blessed or not.
But here’s the rub: Only 29% of respondents trust the accuracy of AI outputs, down significantly from previous years. Nearly half say AI’s answers are “almost right, but not quite,” and 66% report spending more time debugging AI-generated code than they expected. If you’re a platform engineer, that’s a flashing red signal: Uncontrolled AI adoption without guardrails will lead to productivity drains, security gaps and reliability nightmares.
Agentic AI Still in the Sandbox
While generative AI adoption is skyrocketing, agentic AI — autonomous, multi-step AI agents — is still in the early adopter phase. Only 31% of developers use agents today, 17% plan to, and 38% say they don’t plan to use them at all.
For platform engineers, that’s both a challenge and an opportunity. The challenge: Designing platforms that can integrate agentic AI responsibly, with observability, logging and governance baked in from day one. The opportunity: Influence how and when these agents are introduced so they enhance workflows without creating shadow automation that nobody owns.
Claude Rising, ChatGPT Dominant
The survey reveals a shift in AI model preferences. ChatGPT remains the most used tool (82% of devs), but Claude Sonnet is building a strong reputation — ranking high in admiration even if usage still trails. For platform engineering teams, that underscores the need for multi-model support in internal development platforms (IDPs). Developers will want to try new models, and the platform should enable that experimentation without requiring every engineer to go rogue.
The Platform Engineer’s Mandate
This year’s survey data points to a few clear mandates for platform teams:
- Codify AI Best Practices – With trust declining, platform engineers should provide curated AI tools, pre-configured for security, compliance and accuracy verification. Guardrails aren’t optional; they’re the differentiator between chaos and productivity.
- Support the Whole AI Spectrum – From autocomplete to full-blown agents, the platform needs to handle varying maturity levels. Today’s “no agents for me” developer might be tomorrow’s power user — so architect for flexibility.
- Embed Observability for AI – It’s not enough to monitor infrastructure anymore. AI-driven development adds a layer that must be observable: prompts, outputs and downstream impacts. When things go wrong — and they will — debugging without this visibility will be a nightmare.
- Enable Safe Experimentation – The survey shows developers are exploring multiple AI models. A well-designed platform lets them do that in controlled sandboxes, with clear paths from proof-of-concept to production-ready.
- Plan for Cultural Change – Perhaps the most overlooked takeaway: AI adoption is as much a cultural shift as a technical one. Platform engineers must work with dev leads, security and product teams to build trust — not just deploy tools.
In Transition: The Next 24 Months
The Stack Overflow survey paints a picture of developers in flux. Generative AI is embedded, agentic AI is emerging and trust is still catching up. For platform engineers, this is a window of influence. You can shape how AI fits into development at your organization — balancing freedom and governance, speed and safety.
The next two years will likely see agentic AI cross the chasm, and trust metrics rebound — if platform teams get ahead of the curve. Those who wait until agentic AI is mainstream will be playing catch-up in an environment already defined by others.
Final Word: Build for the AI We Have, Prepare for the AI We’ll Get
The 2025 Stack Overflow Developer Survey is more than a snapshot — it’s a blueprint for platform engineering priorities. AI isn’t coming. It’s here. The question is whether your platform will enable developers to use it effectively, safely and trustworthily.
Platform engineers are in the best position to make AI work at scale — not by slowing it down, but by giving it the structure it needs to thrive.