Measured, not marketed.
Every figure on this page comes from a real project run through our own pipeline. We publish what we measured; we don't round up, and we don't publish what we didn't measure.
Four projects, three different stacks — .NET on MS SQL, .NET on a Progress database, and Next.js on Postgres — delivered through one stack-agnostic pipeline, one of them running in production. These are recent deliveries, not the limit: the pipeline builds on the stack your standards require.
A retailer-licensing system — specified, proposed, and built
Our flagship: a web-based retailer licensing application system for a state program. First the specification — the kind that anchors a procurement — produced in days, not months:
Then the delivery: a working system — .NET 8 on MS SQL — built from that specification in three weeks, alongside the full bid pack the same pipeline produced: the proposal, the compliance response, the price sheet, and a clickable demo an evaluation panel can walk through.
A contract-lifecycle-management system — a different stack entirely
The same pipeline delivered a contract lifecycle management system as .NET on a Progress (OpenEdge) database — the kind of stack constraint real agencies actually have. No re-tooling, no exceptions: the specification, prototypes, and build kit adapt to the accepted technology decision, not the other way around.
The pilot that proved the machinery
A deliberately compact contractor-licensing build — three core use cases, Next.js and Postgres — that we used to prove the delivery machinery end to end: document-first intake, guided search, a security assessment, and the full change cycle. It stands today at 487 passing automated tests and a 43/43 end-to-end walkthrough, verified August 19, 2026.
After delivery, the requirements changed — as they always do. A change kit containing only the delta added an entire search module; a reconcile step then verified the delivered code against the updated specification before handover. Small system, full machinery — that was the point.
An event-management system — in production
The first delivered system running live at a client: an event-management platform for a community non-profit — registration, payment declarations with review, slot booking, a cultural programme, expenses, inventory, announcements and reports under a three-role model. Built on Next.js and Postgres and delivered in under three weeks, specification to live; it is taking real registrations and bookings today.
Every use case in the delivered scope was walked through as a clickable prototype before it was built. The specification covers more than was built — on purpose: the client went live on a first release that works, holding a fully specified Phase 2 they can commission when they want it.
The Challenge stage — calibrated against reality
Before any build, an independent AI review pass — the Readiness Review — interrogates every finished use case and produces numbered clarifications: the question to ask, plus the default a builder would otherwise assume. We measured it the only way that matters: against the gaps that actually surfaced after a build.
It found 7.5 out of 10 of them in advance. It fabricated zero findings.
We publish the miss rate alongside the hit rate, because a review you can trust is one that tells you what it is.
Measured projects, not a long track record. The licensing and contract-lifecycle systems are production-ready, built and tested with fictional data, and not yet deployed; the event-management system is in production with real data. Not included in any of these builds, and scoped separately: a software bill of materials and a formal accessibility audit. Deploying into your production environment — real data, live integrations, your security review — is a human-led phase we run with you.
We publish what we measured. We don't round up.