Vendor Manual

A workflow reference for AI tool vendors listing and improving their tools on AECO Shield.

Evaluating tools as a buyer instead? See the Practitioner Manual →

Draft v1 · Sections V.00–V.10 · Methodology v3.4
V.00

How to use this manual

This manual walks the vendor workflow end-to-end — from creating a vendor account through registering a tool, reading its ACS assessment, working the remediation roadmap, submitting evidence for re-scoring, and embedding the Stamp-Safe™ badge on your own site once earned.

Concept definitions live at /faq. Band boundaries and hard-filter rules live at /help/quality-policy. Methodology derivation (why scores are what they are) lives at /methodology. Cross-refs point out to those pages; this manual does not re-derive them.

Who this is for

This manual is written for people whose primary function on the platform is Tool Vendor — the account type that registers AI tools, receives ACS assessments, works remediation, and (once qualified) earns the Stamp-Safe badge for embed.

If your role is evaluating AI tools rather than publishing your own — a Licensed Professional, PM, BIM Manager, or executive at an AEC firm — the Practitioner Manual is the right doc.

V.01

Getting started as a vendor

A vendor account is created at sign-up when you select Tool Vendor as your primary function. Behind the scenes, this creates a row in the vendors table (keyed to your auth id) with your name, email, and company details from the signup form. Your account has no firm — vendors don’t belong to firms (that is a buyer concept).

After signup you land on your vendor dashboard at /vendor-dashboard. The sidebar shows the vendor navigation — deliberately lean, four groups:

Today
Getting started (the milestone hub — see below), Dashboard (your working overview: tiles, tool list, recent assessments), and Checklist (a personal to-do list — free-text or workflow-linked items).
Your Tools
My Tools (all tools you’ve registered, with per-tool verdict + registry-visibility chips), Register a Tool (V.02 below), and Registry (the public /registry — where buyers see your tools alongside the rest).
Reference
Methodology — the ACS vv3.4 specification your tools are scored against. Same doc buyers and insurers read.
Account
Settings — your vendor profile (V.09) and billing (V.09).
V.01.a

The Getting Started hub

The hub at /vendor-dashboard/getting-started is the vendor-appropriate map of the vendor journey. It has two parts: an auto-detected milestone tracker and a launcher grid of the four vendor surfaces.

Register your first AI tool
Self-serve — you complete this by clicking Register a Tool. The other three milestones below are editorial- driven; they complete automatically once the underlying data changes.
AECO scores your tool
AECO editorial runs the ACS methodology (22 questions across 7 layers) against your tool. Auto-detected — no action required from you. See V.03 for what the assessment shows.
Your tool is published to the public registry
After editorial reviews and publishes the assessment, your tool appears at /registry and becomes discoverable to buyers. Auto-detected.
Earn Stamp-Safe™
Requires ACS ≥ 8.8 AND Q21b Liability Transfer ≥ 3. Close roadmap gaps (V.04) and submit evidence (V.05); editorial re- scores; on qualification the badge unlocks (V.06). Auto- detected once earned on any of your tools.

Below the milestones, the launcher grid opens the four vendor surfaces: Register a Tool, My Tools, Registry, and Methodology. The hub is your persistent path back to the capabilities map — bookmarkable as /vendor-dashboard/getting-started.

Editorial cadence

Three of the four milestones (scored, published, Stamp- Safe) are editorial-mediated, not self-serve. This is deliberate — AECO Shield is an editorial platform, not an auto-grader. Your role is to register the tool with accurate details and to work the roadmap once scored; AECO editorial does the scoring and publishing.

V.02

Registering a tool

Go to /vendor-dashboard/tools/new (or click “Register a Tool” from the sidebar or the Getting Started launcher). The form is intentionally short — you supply the identity + basic technical shape; AECO editorial fills in the assessment from public evidence afterward.

Tool name (required)
The public name buyers will see on the registry. Free text. Example: “SiteScan AI”.
Description
Short summary of what the tool does. Optional but strongly recommended — appears on the vendor-facing tool detail page and helps editorial understand scope during scoring.
Version
Optional. If your tool has a version identifier the assessment should be tied to (e.g. 2.1.0), enter it here. Re-assessments on a later version become a new row rather than overwriting.
Tool URL
The tool’s public marketing site. This is the primary source editorial scrapes for evidence during scoring — a rich site with linked docs / ToS / security / privacy pages will score more accurately than a sparse one.
Category
Free-text workflow category (e.g. “BIM Analysis”, “Site Safety”, “Design Generation”). Surfaces on the registry as a filter chip.
Deployment type
Cloud, On-Premise, or Hybrid. Affects some layer scoring (data governance, integration considerations).

On submit, a tools row is created with yourvendor_id set to your auth id (the load-bearing ownership convention every downstream check trusts). You’re redirected to the tool detail page.

What you'll see next: Awaiting assessment

The tool detail page renders an amber “Awaiting assessment” info banner: “AECO editorial will review this tool against the ACS methodology and its Shield assessment — score, layer breakdown, and improvement roadmap — will appear here. No further action is needed from you right now.”

This is the honest state. There is no vendor-triggered scoring — the tool sits at this banner until editorial reviews it. Expect editorial cadence to vary; watch your tool detail for the assessment appearing.

V.03

Reading your assessment

Once editorial has scored your tool, the Assessment Status card on /vendor-dashboard/tools/[id] replaces the “awaiting assessment” banner. Here’s what the fields mean:

ACS score (0–10)
The composite AI Compliance Score. Derived from the 22- question, 7-layer methodology plus S1–S4 safeguards. The full derivation lives at /methodology.
Score band
Textual band: Excellent (≥9.0), Strong (≥7.5), Moderate (≥6.0), Developing (≥4.0), Non-Compliant (<4.0). Cutoffs at /help/quality-policy.
Stamp-Safe classification
Three states: stamp_supporting (fully qualifies), conditional (borderline), or not_stamp_safe. The engine derives this from L1 score, Q21 evidence, S2 penalty on L1, and Q21b tier. The public Stamp-Safe pill on the registry has a stricter gate (see V.06).
Go/No-Go signal
“Approved” (ACS ≥ 7.5), “Conditional” (ACS ≥ 6.0), or “Insufficient for Stamped Work” (ACS < 6.0). This is the fast-signal buyers scan.
V.03.a

The 7 layers

Each ACS score decomposes into 7 layer scores stored in acs_assessments.layer_scores:

  • L1 Responsible Charge — supports the licensed signatory
  • L2 BIM & Digital — technical integration and digital maturity
  • L3 Insurance & Liability — E&O / GL / coverage terms
  • L4 Process & Governance — repeatability, validation
  • L5 Talent & Capability — vendor + end-user competency
  • L6 Adoption & Scaling — deployment maturity
  • L7 Supply Chain — sub-processors, model provenance

The Remediation Roadmap page (V.04) renders these as bars with the weak-line (6.0) marker; the “How you measure up” panel (V.08) adds the top-band display target and the Stamp-Safe L1 gate line.

V.03.b

Additional signals on the tool detail

  • Q21b Liability Transfer banner — appears if the assessment’s Q21b tier is ≤ 2 (concerning or critical). Flags aggressive liability-transfer language in your ToS. Q21b = 1 forces stamp_safe = not_stamp_safe regardless of other scores.
  • OSS version + stability banners — appears for open-source tools with version/conformance risks.
  • Funding tier + investor conflict badges — editorial signals on the vendor’s funding posture.
  • US Market Entry panel — HF1–HF5 hard- filter status (licensure, E&O, GL, DBE) framed for vendors as “here’s what US firms check.”
V.04

Working the Remediation Roadmap

Go to /vendor-dashboard/tools/[id]/remediation (or click “View Remediation Roadmap” on the tool detail). This page has three sections: assessment header, layer scores bar chart, and the roadmap items list.

V.04.a

Where roadmap items come from

Roadmap items are generated automatically by the scoring engine after a re-score, from four sources:

  • Hard filter flags fired — one item per HF1–HF5 that fires (licensure jurisdiction / discipline / E&O / GL / DBE).
  • Weak layers — one item per layer scoring below 6.0 (the Moderate threshold). Severity high if score < 4.0, else medium.
  • Q21b penalty — one item if Q21b tier ≤ 2 (severity critical for tier 1, high for tier 2).
  • S2 critical-response penalties — one item per weight-1.5 critical question scoring < 3.

Items are sorted by severity (critical → high → medium → low), then by layer. Each carries a gap description (what’s missing), a recommendation (how to close it), and an effort estimate (days / weeks / months / quarters).

V.04.b

Item shape

Severity
critical, high, medium, or low. Colored pill on the row.
Layer + question code
Which of the 7 layers the gap belongs to, plus a methodology reference (e.g. Q21b, HF3, L5-DEFICIT).
Gap description
Plain-English statement of what the assessment identified as insufficient (e.g. “L3 scored 5.8 — below the Moderate threshold. Insurance coverage documentation is incomplete”).
Recommendation
What to do to close the gap. Actionable, not aspirational.
Effort estimate
Approximate cadence (days / weeks / months / quarters). Guides prioritization when working through multiple items.

How to prioritize

Start with any critical items — those typically block Stamp-Safe or represent a hard-filter failure. Then work high items in effort- ascending order (weeks before quarters). Closing a layer deficit lifts multiple derived signals: the layer score, the ACS composite, potentially the score band, and potentially Stamp-Safe eligibility if the fix affects L1 or Q21b.

V.05

Submitting evidence for re-score

Each roadmap item has a per-item action affordance — Submit evidence — that opens an inline form. This is how you provide the documentation editorial needs to reconsider the item and, on acceptance, re-score the tool.

V.05.a

The submission form

Evidence type
Auto-detected from the item’s question code, layer, gap description, and clause reference (e.g. an HF3 item defaults to an E&O certificate type). Editable — you can override the dropdown.
Claim summary (required)
A one-line statement of what the evidence proves (e.g. “E&O coverage raised to $2M per occurrence effective 2026-Q3”). ≤ 500 characters.
Evidence URL (required)
A publicly-hosted URL to the supporting document. Must begin with http:// or https://. You host the doc; editorial reviews it in place — no file upload. ≤ 2000 characters.
Note (optional)
Free-text context for editorial — anything a reviewer would want to know that isn’t in the URL. ≤ 2000 characters.
V.05.b

The chip state machine

After you submit, the per-item row shows a status chip that walks the following states:

Submit evidence
Initial state — no submissions yet. The button opens the form.
Evidence submitted — awaiting review
Your submission is queued for editorial. No action needed from you.
Under editorial review
Editorial has started evaluating the evidence.
Accepted — pending re-score
Editorial has accepted the evidence. Re-scoring follows on editorial’s next cycle (not immediate). When the re-score runs, the item may close, the layer score may rise, and the ACS composite may increase.
Rejected
Editorial reviewed the evidence and it did not satisfy the item as stated. Not terminal — you can submit new evidence (a “Submit new evidence” affordance appears alongside the chip).

Honest expectation: editorial mediates the re-score

Submission is not the same as re-scoring. The submission is a record for editorial review — accepting the evidence updates the review status but does not automatically re-score the assessment. Re-scoring happens on editorial’s cycle, after acceptance.

This is deliberate. AECO Shield is an editorial platform (not an auto-grader); the re-score decision lies with editorial, and the vendor’s role in this loop is to provide clean, verifiable evidence that closes the item’s stated gap. When re-scored, the fresh assessment regenerates the roadmap; the accepted item is no longer on it if the gap was closed.

What makes a strong submission

  • Public URL editorial can open without asking for access
  • Dated / versioned document (a certificate expiry, a doc revision date, a page URL that will keep pointing at the same content)
  • Directly answers the item’s ask — not adjacent evidence
  • Claim summary matches what the linked doc actually shows
  • Vendor note explains any indirection (e.g. “section 4.2 of the linked PDF”)
V.06

Earning + embedding the Stamp-Safe badge

Stamp-Safe™ is the public credential your tool earns when it meets the qualification gate. It renders as a pill on your registry entry and as an embeddable badge you can paste on your own marketing site.

V.06.a

The gate

The public Stamp-Safe pill on the registry appears when both of the following hold on your tool’s latest published assessment:

  • acs_final ≥ 8.8
  • q21b_score ≥ 3 (Q21b Liability Transfer Test)

Both gates apply to the assessment currently published to the registry (not to any private draft). Failing either gate shows the “not yet earned” placeholder on your tool detail with a link to the roadmap.

V.06.b

The badge section

When earned, a green-bordered “Your Stamp-Safe badge” section appears on your tool detail page. It contains three parts:

Live badge preview
A 240×80 SVG rendered from /api/badge/[toolId].svg — what viewers see when the embed loads on your site.
Public registry URL
The link a badge click takes viewers to — your tool’s /registry/[toolId] compliance card. Includes a copy button.
Embed HTML snippet
A copy-paste snippet in the shape <a href=“...”><img src=“...svg” width=“240” height=“80”></a> with your production registry URL and badge URL baked in. Paste it into your marketing site’s HTML.
V.06.c

How the badge behaves once embedded

  • The SVG is self-contained — inline system font stack, no external CSS/image references. Renders on any embedding site without follow-up requests.
  • 1-hour CDN cache — the endpoint returns Cache-Control: public, max-age=3600, s-maxage=3600 so hotlinked embeds are absorbed by Vercel’s CDN rather than hitting the origin.
  • If your tool later drops below the Stamp-Safe gate (a re-score fails the ≥ 8.8 threshold, or Q21b tier drops below 3), the badge endpoint starts returning 404 on the next cache expiry — the badge on embedding sites goes blank. The credential is live-derived, not a persistent artifact.

If you haven't earned Stamp-Safe yet

The tool detail shows a placeholder card in place of the badge section: “Stamp-Safe requires an ACS score of 8.8 or higher AND a Q21b Liability Transfer Test tier of 3 or higher. Close your roadmap gaps to lift your score and clear the liability transfer test — editorial publishes the badge after re-assessment.” Open your remediation roadmap (V.04) and work the highest-severity items; submit evidence (V.05) as you close each; wait for editorial re-score.

V.07

How buyers see your tool

Once your tool is published to the registry, a “How buyers see this tool” section appears on /vendor-dashboard/tools/[id]. It gives you direct entry points to the exact pages a buyer / insurer / customer sees when they encounter your tool on the platform.

Open your public compliance card
Links to /registry/[toolId] — the shareable per-tool card that buyers screenshot, insurers reference, and customers link to. Score, band, Stamp-Safe pill, Go/No-Go, and share controls. Opens in a new tab so you keep your dashboard context.
Open the buyer's full report
Links to /assessments/[assessmentId]/buyer-report — the auth-gated, print-optimized 9-section buyer report: executive summary, ACS + band, layer-by-layer breakdown, Stamp-Safe classification, Go/No-Go, insurance exposure, hard filter status, gap report, and remediation roadmap. This is what a buyer prints and hands to their board or E&O carrier.

The section renders only when your latest assessment is registry-public (a private draft doesn’t have anything to link to). No button ever points at a non-public assessment — the surface reflects what actually exists in the buyer world.

Cross-ref

The buyer’s reading of the score screen is documented in Practitioner Manual §03. Skimming that section shows you exactly what a buyer scans on your tool’s card — useful when deciding which roadmap gaps to prioritize for buyer-facing impact.

V.08

How you measure up

The “How you measure up” section on your tool detail places your tool against the methodology’s fixed thresholds — and, as the registry grows, against other published tools.

V.08.a

Standards-anchored view (always shown)

Three parts:

ACS on the scale
A horizontal 0–10 scale with reference ticks at 4.0, 6.0, 7.5,8.8 (Stamp-Safe cutoff), 9.0. Your marker is colored by DCS (green ≥ 8.8, amber ≥ 6.8, red below). Delta text tells you either “+X above the Stamp-Safe cutoff” or “X below Stamp- Safe — need +X to reach it.”
Stamp-Safe gates checklist
Public-registry gates as a pass/fail list: ACS ≥ 8.8 (✓ or ✗) and Q21b Liability Transfer ≥ 3 (✓ or ✗). Shows your actual values alongside. If your engine classification is stamp_supporting, that line appears underneath.
7-layer targets
For each layer, a bar with reference lines: the 6.0 weak threshold (real methodology), a 9.0 top-band marker (labeled a display target — not a formal methodology threshold), and on L1 the 7.0 Stamp-Safe gate line (real methodology). Per-layer delta text — “✓ Top band”, “0.3 to top band”, “0.8 to Strong”, or “weak — see roadmap” — plus a per-row “→ roadmap” link if that layer has an open remediation item.
V.08.b

Peer comparison (conditional)

Below the standards view, a peer-comparison block appears when the registry has at least 8 published, non-yours tools (the pool excludes your own tools so you’re not compared against yourself). It shows your ACS vs the registry median, your percentile, and — when at least 8 pool tools have varied per-layer scores — a peer-mean tick on each layer bar.

If the pool is below the threshold, the section is absent (not empty-state text — the standards view stands alone). The block auto-activates as the registry grows.

Reading conventions

The 6.0 and 7.0 lines are real methodology thresholds — the weak-layer trigger and the L1 Stamp-Safe gate respectively. The 9.0 line is a display target: the methodology’s Excellent band applies to composite ACS, but layering it on individual bars is a display convenience so a vendor has a target to aim at. Not a formal per-layer methodology threshold.

V.09

Your account: settings + billing

The vendor Settings surface at /vendor-dashboard/settings has two tabs: Profile and Billing. Both are session-scoped — you can only edit your own vendor row.

V.09.a

Profile

Your name (required)
Contact person on the vendor account. Shown in greetings and on editorial correspondence about your tools.
Email
Locked to your signed-in account email. Not editable via this form — changing account email requires an auth flow.
Company name (required)
Your canonical company name. Important: changing this applies to new assessments only. Existing published assessments keep the company name they carried at scoring time (they store a tool_vendor snapshot for evidentiary stability — historical records are not re-stamped).
Company website / size / target market / country
Optional fields. Target market is multi-select (federal / private commercial / state & local public). Company size uses fixed brackets (1-10 / 11-50 / 51-200 / 201-500 / 500+).
V.09.b

Billing

/vendor-dashboard/settings/billing shows a read-only plan snapshot. Billing is closed while AECO.digital is in research and preview — there is no payment method on file and nothing to manage.

Vendor plan
Not currently offered — AECO.digital is in research and preview. What a listing covers: public registry listing on qualification, ACS methodology assessment, and (on qualification) the Stamp-Safe badge for embed.
Current plan card
Plan (Vendor / No active plan), Status (Active / Trial / Cancelled / Past due / etc.), and — for active plans — Renews or Trial-ends or Access-ends date.
Manage billing
Not available — billing is closed while the project is in research and preview. No payment method is held and there is no subscription to manage.
V.10

Finding your work

Three surfaces to reach your tools and their state:

Vendor Dashboard (/vendor-dashboard)
Four tiles at the top: Total Audits (published + public assessments on your tools), Active Assessments (tools registered but not yet published — awaiting or in editorial), Stamp-Safe Tools (count with stamp_supporting classification), and ACS Average. Below the tiles, the “Your AI Tools” list shows per-tool chips (verdict + registry visibility) alongside the tool metadata. Below that, the “Recent Assessments” table lists your published rows in reverse-chronological order.
My Tools (/vendor-dashboard/tools)
The fuller table — every tool you’ve registered. Adds an open-gaps chip per row + “Roadmap →” deep-links to the remediation page + a “Detail →” link to the tool detail page. This is the primary working surface once you have multiple tools registered.
Checklist (/vendor-dashboard/checklist)
Your personal working to-do list. Pick a vendor capability from the dropdown to link a “Start →” deep-link, or type any free-text item. Per-user — only you see yours. Persists across sessions.

Getting Started as a persistent path

/vendor-dashboard/getting-started stays useful even after you’ve earned Stamp-Safe — it collapses the completed milestones to a summary line (“Setup complete ✓”) and keeps the launcher grid as a persistent map of the four vendor surfaces. If you register a second tool, the milestones re-evaluate (they’re per-vendor, not per-tool).