5 prompting techniques every platform team should standardize before their next AI rollout

We studied developer experience and productivity data from thousands of engineers actively using AI in their daily work to understand what separates successful rollouts from struggling ones. That research became our Guide to AI Assisted Engineering - a practical, research-backed resource covering the prompting techniques and use cases that actually move the needle when standardized at the platform level.

In this session, Justin Reock, Deputy CTO at DX, will demonstrate each technique with code and prompting examples - with a specific lens on how platform teams can embed these into their toolchain, developer portals, and onboarding flows rather than leaving them to individual discretion.

Understand which prompting techniques are worth standardizing across your platform - and how to actually do it
See examples of meta-prompting, system prompts, and multi-shot prompting applied in the context of platform engineering workflows
Learn where AI rollouts tend to break down at the platform level, and what your team can do about it before the next rollout
Leave with a concrete shortlist of use cases - stack trace analysis, refactoring, auto-documentation - worth building into your platform's developer defaults

Agenda:
Introduction: Why AI rollouts succeed or fail at the platform level; research methodology and high-level findings
Prompting techniques: Meta-prompting, prompt-chaining/recursive prompting, one-shot and few-shot vs. zero-shot prompting — with examples relevant to platform team workflows and toolchain integration
Recommended use cases: Stack trace analysis, refactoring existing code, mid-loop code generation, auto-documentation — and how to standardize these as platform defaults
Closing: Advice for platform engineering leaders on driving adoption, measuring impact, and building prompting standards into your platform before your next rollout