The Delivery Model
Our delivery model isn't a methodology to follow — it's an operating model we install inside your organisation. It brings order, clarity, and predictability to complex delivery environments.
Principles
These aren't slogans. They're load-bearing decisions that shape every engagement.
AI writes the code, the tests, the documentation. Humans provide direction, review, and approval. This isn't about replacing people — it's about amplifying what a small, focused team can achieve.
Teams own products continuously. There's no "build it and hand it off." The people who build it are the people who operate it, learn from it, and improve it.
Projects don't end — they graduate. The investment ratio shifts from building to operating, but ownership remains with the team. Continuity is the default.
Every decision, risk, and design choice is traceable. Artefacts flow bidirectionally through phases — what we learn downstream updates what we decided upstream.
We start with strong opinions on everything from testing strategy to encryption to CI gates. You can override any default — but you need a decision record to do it.
A four-tier model defines when AI acts autonomously, when it recommends and proceeds, when it pauses for human input, and when it blocks entirely. No ambiguity.
The Five Phases
Every initiative moves through five phases. Each has clear inputs, outputs, quality bars, and governance checkpoints. The cycle is continuous — insights from operations feed back into the next iteration.
Establish shared clarity on the "why" at both the product and initiative level. Produce the product brief — the single source of truth for vision, users, outcomes, and constraints. This is where alignment happens, before a line of code is written.
Translate the "why" and "for whom" into "what we need to build." Six research streams run in parallel: users, domain, technical feasibility, risks, competitive landscape, and compliance. For brownfield work, this includes a baseline assessment of the existing system.
Turn discovery into a concrete delivery plan. The shaped pitch defines increments, technical approach, threat model, migration strategy (if applicable), and mockups. Every decision is recorded. Every risk is registered.
AI builds the system in iterative cycles — each shaped increment goes through implementation planning, code generation with tests and documentation, pull request review, and deployment. Artefacts are updated continuously. Human review gates are non-negotiable.
The team shifts from building to running — but never stops learning. Incident management, SLOs, capacity planning, and continuous discovery all feed back into the product brief. When it's time, the system either receives re-investment or begins a governed retirement.
Digital Twins
Digital twins are AI agents that embody professional roles — security architect, UX researcher, compliance officer, and more. They provide structured challenge throughout the lifecycle, ensuring no perspective is missed.
Threat modelling, vulnerability assessment, security-by-design review
User advocacy, journey validation, accessibility review
Regulatory alignment, data governance, audit readiness
Infrastructure review, scalability, operational readiness
Test strategy, coverage analysis, quality gates
Business logic validation, domain model integrity
Data architecture, pipeline design, storage strategy
Load testing, performance budgets, optimisation
Documentation quality, developer experience, runbooks
Metrics design, outcome measurement, insight synthesis
One model, three contexts
The same delivery model adapts to your context. Whether you're building something new, transforming something existing, or iterating on a live product, the phases, artefacts, and governance remain consistent — with context-specific activities where needed.
New systems built from the ground up. Full lifecycle engagement from Envision through to Operate & Learn.
Existing systems that need modernisation. Includes baseline assessment, migration strategy, and incremental transformation.
Live products receiving continuous investment. Streamlined phases focused on discovery, shaping, and delivery of new increments.
We'd be happy to walk you through how the delivery model would apply to your specific situation.
Start a conversation