Sourced fact
What the public record says.
A claim tied to a named source, date, and scope. Unknown stays unknown.
Evidence and analysis method
Show the source. Mark the judgement. Keep the gap.
Anvil North MINT analysis is useful only when a reader can see where the public record ends, where interpretation begins, and what evidence would change the view.
Information classes
Three different statements need three different treatments.
No automated summary, score, or confident tone is allowed to erase this boundary.
MINT assessment
What the evidence may mean.
Anvil North’s reasoned interpretation, with factors, confidence, and unresolved gates shown.
Forecast
What to watch next.
A time-bounded expectation linked to observable indicators and revised as facts change.
Source to publication
A reviewed claim earns its place.
Discovery can begin with automated collection, but publication depends on a separate claim-level evidence and editorial review. Records that do not clear that gate remain internal.
- 01
Capture
Store the source, publisher, URL, date, document version, and access conditions.
- 02
Resolve
Connect names and aliases to the correct organization, capability, program, or buyer.
- 03
Extract
Turn a passage into a bounded claim tied to a specific field.
- 04
Review
Check support, conflicts, classification, uncertainty, and publication fitness.
- 05
Publish
Expose the claim, source boundary, reviewer date, and revision history together.
No opaque master score
Four measures. One gate layer.
Each dimension answers a different question. Combining them into one number would hide the reason a pursuit should proceed, pause, or stop.
- Evidence confidence
- How strongly does the public record support the claim? Source measure
- Capability fit
- How closely does a specific offering map to a stated requirement? Match measure
- Opportunity maturity
- Where is the buyer in the policy, program, and procurement lifecycle? Timing measure
- Market attractiveness
- Is this market worth pursuing for this capability through this route? Decision measure
- Access and compliance
- Can the supplier lawfully and practically participate? Hard gate
Publication rules
What the platform will do.
- Prefer first-party and durable official sources.
- Cite the field or relationship the source supports.
- Retain conflicts and superseded facts in history.
- Show review dates, coverage notes, and explicit nulls.
- Keep deterministic search when AI is unavailable.
- Offer correction and company-claim routes through review.
Safety boundary
What the platform will not do.
- Publish scraped personal contact information.
- Treat an AI answer as the system of record.
- Infer a procurement, partnership, or endorsement without evidence.
- Turn unknown maturity, capacity, clearance, or export status into a positive claim.
- Let paid participation alter analytical ranking.
- Replace legal, security, export-control, or bid advice.
Staging data boundary
This is a product-validation set.
The current site contains 9 full company profiles, 155 compact entries, 1 historical record, and 5 program dossiers. Every public company record has an editorial publication decision, a supported Canadian footprint, a specific technical activity, and direct sources.
The imported source set contained 822 normalized candidates. After entity resolution, scope review, and the publication gate, 666 unresolved or merged records remain internal. This includes 164 associations, ecosystem bodies, student and general flight-training organizations, media and event properties, recruiters, non-technical intermediaries, public bodies, airports, and charities. Source observations remain preserved for audit.
Official-site material establishes what a company says about its own portfolio; it is not a verified-supplier designation or independent performance assessment. Compact entries remain deliberately narrow, and omission from the public directory is not evidence that a capability is unavailable in Canada.