The home for the Advanced Technologies Division's documentation and interactive prototypes — gathered in one place to share across the team, with leadership, and with customers. Some projects build on each other; others stand alone.
The Build Safely offering set behind one entry — two interactive governance demos (the governance-layer simulator and the CMMC L2 readiness walkthrough) plus the service definitions & sample SOWs.
The Night Shift operating model behind one entry — the proposal, the implementation guide, and the leadership-impact briefing on unmanned, bounded-autonomy overnight CI/CD.
The Strong Industries pitch behind one entry — the animated capability showcase (plus a printable leave-behind) and two iterations of the AI Command Center concept.
The commercial overview of what Airiam builds — RAG assistants, integrations, bounded agents, automated workflows and full platforms — and the one thing that sets it apart: everything is governed, controlled, observable and traceable by design, built to support the SOC 2 and HIPAA obligations our clients answer to.
The whole system on one page: eight functional modules over eleven shared services. Defines the information model for business operations and shows how every layer is purpose-built around the governed write-path.
Canopy’s OSCAR platform on one page — the dental/DSO-focused sibling of the AIP architecture: the shared OCRM model and governed write-path, with dedicated customizations and per-deployment integration seams. Under active development.
OSCAR's Plane 2 observability fabric — a swappable, in-process layer over OpenTelemetry that captures metrics, traces and cost from every service and redacts at a single boundary so telemetry carries no PHI by construction. Walks the emit → redact → store → aggregate → read pipeline and the target topology, marking what is live, merged, and proposed across three increments.
Watch a single action walk the eleven-step governance chain — risk-scored, policy-checked, and written to the immutable trail. Pick a scenario and see how the same path lands a different verdict.
One assistant, many hats, provably controlled. Two planes share one seam: a read side where an Authority Context and soft wall keep client and IP data from bleeding between engagements, and a write side where every mutation is risk-scored, policy-checked, and gated to a verdict.
The centralized operator hub for both the AIP and OSCAR platforms (brand-toggle; agnostic core with per-platform specifics): busy-hour analysis, Erlang B/C capacity, cost forecasting, top problem signatures, AI trust services, and drift signals — backed by a full audit view.
A status report (updated 25 Aug 2026) on wiring live OSCAR telemetry into the Ops Console through a governed read seam — the four-hop, three-identity auth path it required, the platform auth hardening it uncovered, and an honest merged-vs-deployed, per-surface “live” accounting across two repositories.
From any source to a triaged Linear issue. Phase 1 is the async pipeline — PII redaction, AI triage, deduplication, deterministic priority. Phase 2 puts six conversational channels (voice, SMS, WhatsApp, chat, Teams, Slack) out front, with warm human handoff.
Embezzlement and fraud detection. Every financial event is scored against eighteen deterministic rules and each actor's own learned baseline. Four bounded agents run the loop end to end — and money never moves without a human at the irreversibility points.
A single-page brief on Airiam Sentinel — the AI controller that learns each person's normal. Every money-moving action is checked against eighteen deterministic fraud controls and the actor's own baseline, then waved through, held for a glance, held for sign-off, or stopped cold. No black box; every flag is explainable.
A single-page brief on Wave 10 of the Design Desk assistant: it now produces phased project plans and level-of-effort estimates even with no comparable internal document, grounding Airiam-specific facts in retrieved documents and clearly labeling everything it assumes. Admins can edit its standing directions live under Prompt Sections.
Four capabilities through one architecture model: plain language to an Architecture IR, to a diagram, to a validated discrete-event simulation, to runnable FastAPI — then production traces reconciled back against the IR. The closed loop nobody else has connected.
A simulation of AI security and governance analysis — how risk surfaces are assessed and controls are reasoned about across the platform.
A regulated-client inquiry reframed as a productized AI enablement & governance engagement — arcing from a fixed-fee assessment through platform enablement and a governed sandbox into recurring managed AI operations, positioning Airiam as the client's standing AI governance function.
An SMB-focused AI business briefing and market-engagement plan — the audience and message, a recommended customer engagement model, and the Night Shift plan framed for market positioning.
A CPA-led, education-first growth channel with two clean fulfillment lanes — Airiam for non-dental SMBs, Canopy / VELLA for dental and DSOs — converting warm, advisor-sourced risk reviews into recurring revenue. Recommends a gated 90-day pilot; models ~3.4× revenue-to-cost over three years.
Internal review of the Olsen Palmer “Institution Profile Automation” project — where the build stands and how OP’s process works across its two integrations. Prepared for Kuk.
The weekly ATD progress review for CEO Kuk Yi (Jul 14, 2026) — Eisenhower priority matrix, this week’s key wins, the platform roadmap, and completed projects.