Systems Development Framework

Visioning Lab has developed a Systems Development framework to help organisations work through complex problems in a structured, transparent way.

The approach moves from defining inputs through to action, enabling clearer thinking, better decisions and ongoing system insight.

Framework Overview

Define → Structure → Insight → Action

Define

Capture and clarify what matters across multiple sources and perspectives.

• Documents, meetings, data, stakeholder inputs
• Structured into clear points
• Identification of assumptions and implications

Structure

Organise relationships, options and decision logic.

• Linking points into models
• Defining options and trade-offs
• Making reasoning explicit and comparable

Insight

Identify patterns, signals and system-level implications.

• Recurring issues across inputs
• Hidden assumptions
• Emerging risks and opportunities

Action

Translate insight into interventions and next steps.

• Decisions and action plans
• Policy or operational changes
• Feedback into future cycles

What this enables

• Structured understanding of complex systems
• Transparent and explainable decision-making
• Comparability across cases and organisations
• Continuous learning over time

Interactive Tool

An interactive Next.js implementation of the framework lives in this repo. It walks the user through Define → Structure → Insight → Action, with LLM-assisted extraction at each step and editable outputs throughout.

Setup

npm install
cp .env.example .env.local   # add your ANTHROPIC_API_KEY
npm run dev

Open http://localhost:3000.

Stack

• Next.js 14 (App Router) + TypeScript + Tailwind
• Anthropic Claude (claude-sonnet-4-6 by default) for extraction & generation
• zod schemas for the Point / Option / Insight / Action data model
• jspdf for the PDF export
• Session state autosaved to localStorage (no server storage)

Project layout

• app/page.tsx — landing
• app/tool/ — the four-step flow + export page
• app/api/ — LLM endpoints (one per step)
• lib/schema.ts — data model
• lib/prompts.ts — system prompts per step
• lib/anthropic.ts — Claude client + JSON helper
• lib/session.tsx — React context + localStorage autosave
• lib/pdf.ts — PDF + JSON export

Applications

• Infrastructure and environmental systems (e.g. water, energy)
• Policy and regulatory decision-making
• Research and interdisciplinary projects
• Complex programme design

Relationship to Existing Work

This project builds on Visioning Lab’s previous work, including:

• Ontology Maker (structured terminology and relationships)
• Water Decisions (decision support for infrastructure systems)

It brings these together into a single, integrated framework.