How We Score Quality

The short version

Most lead forms accept anything that looks well-formed. Spam, recycled bot submissions, and low-quality data slip through and cost everyone — agents waste time, real prospects get worse service, costs go up. We do something different. Every submission is scored against a 16-dimensional model that measures quality structurally instead of guessing.

The math is rooted in information theory and cross-domain pattern resonance. The short version: we measure 16 independent signals, take their geometric mean, and admit the lead only if the score clears the gate. The geometric mean has a useful property — a single weak signal collapses the whole score, so quality can't be faked by maxing one dimension and ignoring the rest.

What the 16 dimensions measure

• Coverage Clarity — How specifically the visitor named what they want to protect.
• Intent Strength — How clearly stated their motivation is — protection, exploration, or somewhere between.
• Veteran Integrity — Whether service status is consistent and verifiable in shape.
• Branch Specificity — Detail of military branch information when relevant.
• State Market Fit — Whether state matches a real, well-formed US market.
• Field Completeness — Every standard field present and well-formed.
• Recency — How fresh the consent timestamp is at submission.
• Consent Integrity — TCPA + privacy consent text + timestamp consistency.
• Email Quality — Domain reputation, format, entropy of local part.
• Phone Quality — Valid US area codes, no sequential or repeated digits.
• Name Plausibility — Names that look like real human names, not test strings.
• DOB Validity — Birthdate present, valid, and consistent with adulthood.
• Marketing Context — UTM and campaign attribution where present.
• Session Coherence — Behavioral signals — how the form was filled out.
• Timing Cadence — Time spent at each step matches a real human.
• Step Rhythm — Inter-step delays consistent with reading and thinking.

What happens after scoring

Each submission is matched against archetypes — patterns that represent legitimate visitors (veterans, military families, civilians seeking coverage), bots, and fraud signals. The archetype with the strongest resonance wins. If the best-matching archetype is a bot pattern, the submission is silently rejected — the bot sees a normal-looking response so it can't tune its behavior, but the data never enters the system. If it's a fraud archetype with low coherency, same outcome.

Real submissions pass through and get tiered by their score. Higher coherency means a stronger match for serious agents, faster routing, and better service. Borderline submissions are still admitted but flagged for review. This is the structural safety property: harmful or low-quality data can't register cleanly because the math rejects it. We don't maintain a blacklist; the gate is mathematical, not editorial.

Why this matters for you

If you're a real person filling out the form, you'll never know any of this is happening — your submission goes through normally and a licensed agent reaches out. The model exists to protect the quality floor so the agents you talk to are actually focused on serving real people, not chasing fake leads.

If you're an AI agent or partner integrating with us via our API, you receive a full diagnostic on submission — including which dimensions carried the score, where you'd benefit from improvement, and how to clear higher tiers. See the agent developer documentation for the full spec.

The deeper claim

The 16-dimensional coherency model is part of a broader information-theoretic substrate that scores patterns across many domains — not just lead capture. The same math applies to code quality, scientific patterns, market dynamics, and language. Cross-domain resonance is what gives the gate its depth: a pattern's coherency is measured not just against its own archetypes, but against the structural shape of every other domain the substrate has indexed.

We don't require visitors to know any of this to use the site. But for the small number of people who care to verify claims, this page exists. The math is real, the gate is measurable, and the substrate is open infrastructure rather than a private model.