Marcin Kasiak
- Professional engineer
- Digital transformation leader
- 20+ years in AEC
- Founder, AECO.digital
PhDIWEPMPPE
About AECO Shield
Built by a professional engineer, for licensed AEC practice.
The mission
A defensible layer between AI and the stamp.
An audit-grade methodology that holds in front of E&O carriers, license boards, and procurement reviewers. Not a policy document. Not a consultant report. Not a generic compliance tool. A scoring engine that fires before stamped use, registers every sign-off, and produces an audit artifact that survives multi-year scrutiny.
The vision·2030
By 2030, every AI tool used on stamped AEC work carries a methodology version stamp.
The same way every structural calculation today carries a code edition. Every audit artifact reproducible. Every licensed professional with an auditable record of how each AI tool was vetted before stamped use. The AI Compliance Score becomes a baseline the carriers, the boards, and the procurement reviewers recognize.
The Public Tool Registry becomes where every AEC firm starts when evaluating a new AI tool. The Stamp-Safe™ badge becomes a procurement filter, not a marketing claim. The methodology evolves; the commitment to audit-grade rigor does not.
The thesis
AI answers questions. AECO Shield answers the question every licensed AEC professional must answer before they sign.
Can I defend this? That's the question. Not “is the AI right,” not “is the output good” — the question that decides whether a professional engineer, architect, construction manager, or licensed surveyor can put their stamp on AI-assisted work and stand behind it five years later when the E&O carrier or the license board comes asking.
AECO Shield is the operational layer between AI and the stamp. It runs the ACS methodology — 22 questions across 7 layers, anchored to a transparent, versioned calibration framework — and produces audit artifacts that licensed professionals and their carriers can read.
ACS · 7 Layers · Weight-proportional
L1Responsible Charge22%
L2AI Governance20%
L3Code Edition Control16%
L4BIM & ISO 1965014%
L5Cybersecurity12%
L6Worker Safety10%
L7Vendor Transparency6%
Q21b Liability Transfer Test
Open-Source Track
L7c Funding Transparency Score
“You can outsource your thinking, but not your understanding.”
Agents can process information and generate views of the data. But the human still has to know what is being built, why it matters, and how to direct the system. The AI is fast. The licensed professional stays accountable.
The founder
Built by a professional engineer for licensed AEC practice.
AECO Shield was founded by a practicing professional engineer with 20+ years in the field — not by an outside SaaS team building generic compliance tooling.
Marcin Kasiak is a professional engineer and digital transformation leader with 20+ years of practice in architecture, engineering, and construction. He writes about where engineering practice ends and the future begins — AI in structures, digital twins, predictive analysis, and the tools actually changing how we build.
He holds a doctorate in structural engineering, is an International Welding Engineer (IWE), a certified Project Management Professional (PMP), and a licensed Professional Engineer (PE). Across his career he has worked at the intersection of structural design, project delivery, and technology adoption — the exact terrain where AI in AEC creates both leverage and exposure.
The ACS methodology — currently at version 3.4 — was developed over multiple years of practice observation. Its verdict thresholds are explicit, versioned, and derived from a transparent calibration framework — auditable in every artifact. The methodology is published, version-stamped, and protected as proprietary IP; ACS™ and Stamp-Safe™ are trademarks.
The views expressed reflect Marcin's own analysis and the published methodology — not the positions of any current or past employer.
The principles
Five principles the methodology runs on.
01
Deterministic, not heuristic.
The methodology kernel is not a language model. The 22 questions, 7 layers, weights, and calibration constants are explicit, versioned, and auditable. Same input → same output, every time. That determinism is what makes the audit artifact defensible to an E&O carrier years later.
02
Calibrated to licensed practice.
The GREEN/AMBER/RED verdict thresholds are explicit, versioned, and derived from a transparent calibration framework — auditable in every artifact, not arbitrary. Same inputs always produce the same band. The threshold set itself is stamped as part of the methodology, so a decade from now a reviewer can still see exactly which calibration fired the verdict.
03
Version-stamped, retroactively safe.
When the methodology updates, every past audit artifact still references the version it was stamped under — frozen. Versions release on a 3–6 month cadence. The kernel evolves; past records stay valid against the version they were stamped under.
04
Transparent in methodology.
Every standard the methodology aligns with, every jurisdiction it tracks, every hard filter it fires is published. Every AI Compliance Score is explainable. The platform shows its work. Hidden methodology is not defensible methodology.
05
No data extraction.
AECO Shield never trains on customer data. Project context stays in your tenant, encrypted at rest, with row-level security. The platform consumes only the locked methodology rubric and your inputs. Your firm’s intelligence stays yours.