Six layers. One brand. One domain.

AECO.digital is one integrated system with four outcome surfaces. Each app expresses the AECO Compliance kernel into a different decision for a different buyer. The methodology is the moat — change one compliance rule in the kernel and every app updates simultaneously, because there is only one data layer.

An operating system for AI in AEC — not a feature collection stapled together.

The public marketing surface at aeco.digital. Every route is public — no auth required — readable by procurement officers, license boards, insurance carriers, and prospective firms before anyone signs up.

Home, /shield, /kernel, /methodology, /registry, /pricing, /about, /faq, /contact — all live here. The registry is publicly browsable; every published tool assessment is readable by anyone.

The fixed substrate: where the user is licensed (jurisdiction, board), under what credential (PE / AIA / CCM / LS / other), and the firm's regulatory geometry (federal exposure, DBE requirements, insurance minimums).

Doesn't change between sessions or projects. The rest of the stack reads from it — hard filters fire against it, FitScore weights against it, audit artifacts stamp it.

ISO 42001 · NIST AI RMF · EU AI Act · CMMC 2.0 · ISO 19650 · IBM AI Governance — the upstream standards that any AEC AI decision has to answer to. AECO.digital is the downstream operational layer that binds those frameworks to the day-to-day work.

When a standard updates (a new NIST profile drops, the EU AI Act adds an Annex, a new ISO revision lands), the Standards Watch queue surfaces the change to the curator, and the kernel re-binds against the new source — no downstream app re-write required.

Audit log, public + private registry, firm maturity profile, project AI records, generated artifacts — one Supabase backend, row-level security on every table. Firm-scoped data never crosses firm boundaries; vendor-scoped data never crosses vendor boundaries; only the published registry is public.

Every score, verdict, and artifact references this layer — never re-derives it. That's what makes a five-year-old sign-off receipt still resolvable against the exact methodology version and the exact data the score was calculated from.

Audit records are retained per our published Quality Policy.

22 questions across 7 layers. Verdict thresholds are explicit, versioned, and derived from a transparent calibration framework — auditable in every artifact. Same inputs always produce the same score; the threshold set itself is stamped as part of the methodology.

The kernel is the moat: changing the methodology changes every downstream score simultaneously, because there is only one data layer. Change one compliance rule in the kernel and every one of the four apps updates — because they all read from L3.

AECO Compliance is published and citable — not sold separately; it's the DNA of every app. See the kernel deep-dive or the full methodology.

Each app expresses the kernel into a different decision. One kernel, four outcome surfaces — one for each buyer, one for each verb they need to answer.

The cockpit is a light shell — not a heavy meta-app that replaces its children. It provides the app switcher, cross-app notifications, global search, settings, and the reports register that spans the apps.

Shield's dashboard IS the Shield home. The cockpit doesn't stand in front of it; it sits around it — thin chrome that keeps the platform identity consistent as the visitor moves between apps.

Only Shield is live today, so the cockpit is single-app in practice; the shell is scaffolded to receive Bid / Strategic / Career as they ship without a rewrite.