Skip to content
Agentic Engineering

Planning & Design

Agents accelerate planning by exploring options, surfacing patterns, and drafting specifications—without replacing stakeholder judgment.

Requirements & Planning

Section titled “Requirements & Planning”

Where agents help

Section titled “Where agents help”

Breaking down ambiguity

  • “What questions should we answer before implementing [feature]?”
  • “What edge cases should we consider for [requirement]?”
  • “Break down [epic] into implementable user stories”

Research and exploration

  • “What approaches exist for [problem]? Summarize pros and cons.”
  • “What are common pitfalls when implementing [feature type]?”

Treat this as research assistance, not authoritative answers.

Specification drafting: API contracts, data models, interface definitions, acceptance criteria. These drafts need human refinement, but they accelerate the starting point.

Estimation support: “Based on this spec, what are the major implementation tasks?” Agents decompose work; estimation remains human judgment.

Where agents struggle

Section titled “Where agents struggle”
  • Stakeholder intent — They can’t replace stakeholder conversations
  • Organizational context — Team ownership, historical decisions, constraints
  • Prioritization — They enumerate options, but can’t tell you what matters most

Prompt patterns

Section titled “Prompt patterns”

User story refinement: “Given this requirement: [paste requirement]. Generate user stories in standard format (As a… I want… So that…). Include acceptance criteria for each.”

Risk identification: “We’re planning to implement [feature]. What technical risks should we consider? What could go wrong?”

Architecture & Design

Section titled “Architecture & Design”

What agents offer

Section titled “What agents offer”

Broad pattern knowledge: Common approaches for your problem type, pattern variations and tradeoffs, anti-patterns to avoid. Doesn’t replace experience, but accelerates exploration.

Articulation: Generate diagrams from descriptions, document decisions, create viewpoints for different audiences.

Challenge and critique: “What could go wrong with this design?” “What am I not considering?” They surface considerations you might miss.

What agents can’t do

Section titled “What agents can’t do”
  • Make decisions — They lack context about your team, constraints, and what you’re optimizing for
  • Understand evolution — They see a snapshot, not trajectory (why things are the way they are)
  • Navigate tradeoffs — They enumerate options, not which tradeoff fits your situation

Prompt patterns

Section titled “Prompt patterns”

Design exploration: “I need to design [type of system]. What architectural patterns are commonly used? For each, what are the key tradeoffs?”

Design critique: “Here’s my proposed architecture for [system]: [description]. What potential issues should I consider? What am I missing?”

ADR drafting: “Help me write an ADR for deciding to use [approach] instead of [alternative]. Context: [provide context].”

Diagram generation: “Create a [type] diagram showing [components and relationships]. Use [format, e.g., Mermaid syntax].”

Resources

Section titled “Resources”

Specifications & Planning

Section titled “Specifications & Planning”

Case studies

Section titled “Case studies”