Intent-Driven Development
Where Human Intent Meets Machine Execution
Agentic AI can build almost anything, at extraordinary speed. The question is no longer what we can build. It is whether we are asking for the right thing. IDD provides a practical framework for specifying, measuring, governing, and scaling intent – so that autonomous systems build what you actually meant.
The Shift
Speed without clarity
Organisations are adopting AI tools enthusiastically, generating code at unprecedented speed, and finding themselves no closer to delivering the right outcomes. The bottleneck has moved. It is no longer about what technology can do. It is about how clearly humans can express what they want technology to do.
Speed without clarity isn’t progress. It’s just faster failure.
The Framework
Four connected capabilities for intent-first delivery
IDD provides a complete model, from individual specification through governance, measurement, and organisational design.
01
Specify Intent
02
Govern Execution
Risk Dials give organisations explicit control over how much autonomy AI systems are granted. Start cautious, build trust through measurement, expand where evidence supports it.
03
Measure Alignment
Intent Fidelity measures how closely implementations match intent across four dimensions. The same standard applies to human and AI implementations alike.
04
The Intent Hierarchy structures shared intent at Organisation, Domain, and Project levels. Intent flows down. Learning flows up. Architecture becomes promoted intent.
Govern intent.
Delegate execution.
The ebook
The complete methodology in fourteen chapters
Part I – Foundations: Why intent matters now, and why the framework survives rapid AI evolution.
Part II – The Framework: How to specify intent, maintain human control, measure alignment, and work with multi-agent systems.
Part III – Scaling: How roles evolve, the Intent Hierarchy, and the Maturity Model.
Part IV – The Living System: Intent evolution, the learning organisation, and the future beyond software.
Who is this for
Intent is a language everyone can, and should speak
Engineers & Architects
From implementer to architect of constraint
Product & Delivery Leaders
Intent as the central artefact
CTOs & Senior Leaders
Scale autonomy without losing control
An Example
Worked Example: Abandoned Cart Recovery
This downloadable specification demonstrates how the six elements of an Intent-Driven specification work together in practice, applied to an e-commerce abandoned cart recovery system.
The example is deliberately comprehensive. It includes full intent articulation:
- Domain model definitions with bounded contexts and aggregates
- BDD acceptance criteria, detailed validation scenarios and edge cases
- Performance benchmarks
- Compliance requirements across GDPR, PCI-DSS, WCAG, and SOC 2
- A structured ethical impact assessment with mitigation strategies
Not every specification needs this level of detail, but teams new to IDD will find it useful as a reference for what a thorough specification looks like when the stakes are high.
Use it as a template for your own specifications, as a training tool to walk your team through what intent-first thinking produces, or as a benchmark to identify gaps in your current approach.
This is the first of what will become a growing library of worked examples. As IDD matures and practitioners begin applying the methodology across different domains, new examples will be added, covering different industries, complexity levels, and specification depths.
If you have a worked example you would like to contribute, get in touch.
About
A living methodology
The ebook is version 0.2, first published on 13th April 2026.
IDD is a living methodology. As AI capabilities mature, as more organisations begin to practise intent-first development, and as real-world case studies emerge, the framework will evolve.
Future versions will deepen the measurement frameworks, introduce tooling that supports intent specification, share patterns from early adopters, and explore Human Intent as a discipline that extends beyond software into leadership, strategy, and organisational design. This document will grow alongside the discipline it describes.
Created by Richard Stockley
Fractional IT Architect, Executive Coach & Mentor with almost thirty years of global experience across software engineering, IT architecture, and delivery leadership. Fellow of the British Computer Society (FBCS). EMCC-accredited Coach and Mentor.
Intent-Driven Development grew from the intersection of consulting on how organisations build technology, coaching leaders on how they make decisions, and a conviction that clarity of intent determines the quality of outcomes.