methodology-process

Measuring whether the AI actually helped users accomplish their goals.

Interpreting implicit and explicit feedback — edits, regenerations, abandonment.

Designing around token limits, memory, and conversation persistence.

Turn-taking, repair sequences, grounding, and dialogue structure for human-AI interaction.

User correction, thumbs up/down, inline editing, and reinforcement signals.

Reading user emotional state from text signals — caps, punctuation density, repetition, latency — and adapting before the user disengages.

Designing interfaces where AI generates UI components dynamically.

When the AI leads vs. when the user leads, and how to hand off control.

Coordinating text, image, voice, and tool-use modalities in a single interaction.

Revealing AI capability gradually to match user mental models.

Designing reasoning chains that produce better outputs.

Defining output format, length, tone, and content boundaries within prompts.

Designing what information goes into the context window and in what order.

Crafting examples that steer AI behavior effectively.

Managing prompt iterations, testing changes, and tracking what works.

Anatomy of effective system prompts — role, context, constraints, format.

Creating reusable, parameterised prompt templates for consistent outputs.

Ensuring the AI behaves predictably across sessions, edge cases, and modalities.

Adapting AI behavior for different cultural contexts, languages, and norms.

Tailoring AI behavior for specific professional domains.

How the AI responds to user frustration, confusion, delight, and distress.

How the AI communicates mistakes, uncertainty, and limitations gracefully.

Defining AI character, voice, and personality traits.

Adjusting formality, warmth, confidence, and style per context.

Capture architectural decisions made during Claude Code sessions as structured ADRs. Auto-detects decision moments, records context, alternatives considered, and rationale. Maintains an ADR log so future developers understand why the codebase is shaped the way it is.