Score AI tools against the methodology, gate them per project, and store the audit trail a license board or E&O carrier will accept. App 1 of 4 on the AECO.digital platform.
Methodology v3.4 · ACS™ · Stamp-Safe™
Inside AECO Shield
Each module reads from the AECO Compliance kernel. Change one methodology rule and every module updates simultaneously — because there is only one data layer.
Score any AI tool against the 22-question ACS rubric across 7 layers. Hard filters HF1–HF5 (licensure jurisdiction & discipline, E&O, GL, DBE) block non-compliant tools before procurement. Stamp-Safe classification per tool.
Outputs
Firm Maturity Profile across D1–D7 dimensions with peer benchmarks and a maturity radar. Establish the firm baseline your FitScore calculation depends on.
Outputs
Rank AI tools against your firm profile and project context — a FitScore 0–10 per tool with hard-filter blocks flagged. Ranked, defensible tool selection for procurement.
Outputs
Sign off on tool use per project with password + identity attestation (four attribution methods). Generates a Project AI Register and AI Liability Audit Report. Log shadow-AI use via a reporting form.
Outputs
Public registry of assessed tools. Vendors claim their tool, get a remediation roadmap, submit evidence for re-score, and earn an embeddable Stamp-Safe badge.
Outputs
What this actually looks like
Not a sales pitch. A walk-through of one tool, one project, one licensed professional, one sign-off. End-to-end in five steps.
Illustrative walkthrough — a structural PE evaluating an enterprise AI assistant for a hospital seismic-retrofit project.
Step 1 · C1 · Scan
Run the candidate AI tool through the 22-question ACS rubric. Hard filters HF1–HF5 evaluate first.
Output · ACS score with band and Stamp-Safe verdict; hard-filter flags surfaced if any fire.
Step 2 · C3 · Compare
Add competitor AI assistants to the compare view. FitScore ranks each against the firm profile plus project context. Example rank: Enterprise Assistant A · 7.2 · Enterprise Assistant B · 6.4 · Enterprise Assistant C · 4.1 (HF3 block).
Output · Ranked FitScore list with block reasons per tool.
Illustrative — not actual assessment scores.
Step 3 · C2 · Maturity
The Firm Maturity Profile establishes D1–D7 posture and peer benchmark. A weak D3 (AI Readiness) throttles the MaturityMatch component of the FitScore.
Output · D1–D7 maturity radar with peer medians; roadmap items surfaced for weak dimensions.
Step 4 · C4 · Sign-off
On the project's AI Gate, the PE signs off on the tool with password plus identity attestation (four attribution methods available).
Output · Sign-off receipt PDF stamped with methodology v3.4 and attribution timestamp.
Step 5 · Audit
The sign-off, register entry, and any shadow-AI reports roll into the audit trail. Retention: 7 years per our published Quality Policy.
Output · Project AI Register PDF plus AI Liability Audit Report — the artifact your board and E&O carrier can read.
Without Shield: a transcript and a hope. With Shield: an audit artifact the carrier accepts.
Four tracks, one platform
Shield answers the same underlying question through four different lenses — one per licensure track.
“Can I stamp this?”
Shield returns a Stamp-Safe classification per tool: whether responsible charge can be placed on its output. Q21b flags aggressive liability-transfer language in the ToS. The audit artifact documents your decision to your board and your E&O carrier.
“Will my E&O cover this?”
Hard filter HF3 checks whether the tool’s E&O coverage clears the project minimum. HF4 does the same for GL. If either fails, the FitScore never rolls up — the block fires before procurement.
“Which tool wins this RFP?”
C3 ranks candidate AI tools against the firm’s maturity profile and the project’s risk sensitivity. HF5 checks DBE subcontractor capacity on public projects. The ranked list — with block reasons — is your decision record.
“Is this valid in this state?”
Hard filter HF1 checks whether the tool’s licensure jurisdiction covers your project state. HF2 checks discipline alignment. Wrong jurisdiction or wrong discipline — the tool is blocked before it enters the FitScore.
Case library — illustrative scenarios
Illustrative scenarios showing the modules used and the qualitative outcome — anonymised firm names, no real institutions or dollar figures.
Vertex Architecture
Evaluating a design-automation AI for a mixed-use tower in a high-seismic zone.
C1 + C3
Outcome · FitScore below the risk-sensitivity threshold. No-Go decision documented before RFP response.
Harlow Contractors
Bidding a federal medical-facility contract with CMMC 2.0 exposure.
C1 + C4
Outcome · HF3 (E&O) triggered on the incumbent AI stack. Tool swapped before contract award; sign-off filed against the compliant replacement.
Northfield Architecture
Firm-wide review of an AI generative-design suite after a vendor ToS update.
C1 + REG
Outcome · Vendor submitted evidence via the portal. Re-score raised Q21b from 2 to 6; roadmap item cleared. Proceeded with conditions.
Drafthouse Architecture
Practitioner flagged an unauthorised AI-assist plugin used on a stamped drawing set.
C4
Outcome · Reported via the shadow-AI reporting form. Event logged to the audit trail with attribution method and timestamp; remediation memo generated.
Summit Civil Engineering
Multi-state PE firm assessing an infrastructure AI tool across 8 jurisdictions.
C1 + C2 + C3
Outcome · HF1 fired in 3 of the 8 jurisdictions. D1–D7 maturity profile flagged a D3 AI Readiness gap. Tool approved for the 5 clear states with a governance-uplift plan attached.
Pacific Rail Consultants
Renewing E&O with a carrier requesting AI-governance evidence.
C4
Outcome · Firm AI Liability Audit Report (portfolio rollup across active project sign-offs) provided to the carrier. Renewal documented against the audit trail.
Positioning
RegTech ships policy. Shield ships an audit log. The difference shows up when a carrier asks for evidence.
What RegTech ships
What AECO Shield ships
Every artifact carries methodology v3.4. When the methodology updates, past records stay stamped against the version they were signed under.
Get Shield
Pro ($149/user/mo), Firm ($799/mo), or Enterprise (custom). 14-day free trial on Pro and Firm — cancel anytime. The same methodology kernel runs across every tier; tiers differ in scale, seats, reports, and support.