How we build

Architecture first. Numbers computed, not generated.

Five principles that govern every product in the neart ecosystem — deterministic engines for the figures of record, governed AI on top.

Architecture before everything

Nothing gets built until the system is designed. Data models, logic flows, compliance requirements — all specified and reviewed before a single line of production code. This isn't bureaucracy; it's how you avoid building the same thing three times, and how every product ships knowing its domain completely. The knowledge is specified, not improvised.

Computed, not generated

The figures of record — pay, earned value, scores, risk ranges — are produced by deterministic engines with no AI in the calculation path. They run identically with the AI switched off. AI drafts, explains and accelerates the work around them; it never computes the number you would take to a board or an auditor. That line is drawn inside every product, deliberately.

Governance by design

Every system is governed by its specification. Architecture, data models and compliance rules are reviewed and locked before development begins. Nothing is built that wasn't first designed; nothing ships that wasn't first verified. Every sensitive action runs inside a boundary that knows exactly what it is authorised to do — and is recorded when it does it.

Uncompromising at every tier

Enterprise standard is not a pricing tier — it is the baseline every product is built to. Every tier delivers the same architectural rigour, the same compliance standard, the same quality of output. What changes is scale and capacity. What never changes is the standard. The price changes; the quality does not.

Systems scale. Headcount doesn't have to.

Traditional software teams grow by adding people. We grow by building better systems — the architecture carries the complexity, not the size of the team. That is what keeps the whole ecosystem lean, coherent, and held to a single quality standard from one product to the next.

The pipeline

Every product moves through the same five stages.

01

Specify

Data models, logic flows and compliance requirements are written down and reviewed — before any production code.

02

Review

The architecture is challenged and locked. If it cannot be specified, it is not ready to build.

03

Build

AI accelerates execution against the governed specification. It implements the system; it does not design it.

04

Verify

Core calculation logic carries 100% branch coverage. Loading, empty, populated and error states all work — on real data, never mocked.

05

Govern

Every sensitive action runs inside a governed boundary, and every output traces back to the inputs and the logic that produced it.

One

standard at every pricing tier

6

countries of payroll (Mellow)

4

AI search engines covered (Inspected)

Spec-first

design before build, always

In practice

What that looks like in the codebase.

Clear separation

Logic, data and interface layers are always kept separate.

Deterministic operations

Critical calculations run on controlled engines, never probabilistic AI.

Structured data models

Every product starts from a complete, reviewed data schema.

Full traceability

Every output traces back to its inputs and the logic that produced it.

UK-first compliance

GDPR and Employment Rights are built in from the specification, not bolted on later.

Zero-tolerance testing

Core calculation logic carries 100% branch coverage before it ships.

Audit-ready

Every output traces to its source

Deterministic core

Numbers computed, not generated

Spec-governed

Design before build, always

UK-first compliance

Built in from the specification

This is how we build software you can trust with the parts of your business that matter most — the numbers computed, the decisions governed, the work shown.