AI Guardrails

How the LLM layer is scoped, and what it's deliberately not allowed to do.

The short version: the AI layer scores companies, not people — it evaluates firmographic account data against your ICP, never a named individual, never sends anything on its own, and every output it produces is meant to be reviewed by a human before it affects a real account.

Scores accounts, not individuals

Every account gets a transparent, rule-based fit_score first — a plain weighted breakdown across industry, company size, revenue, tech stack, and geography, fully auditable with no LLM involved. The LLM layer adds a qualitative account_fit_score and rationale on top, meant to surface context the rules can't (recent funding, hiring surges, competitor mentions) — not to replace the rule-based number or make a final call on its own. The prompt sent to the LLM deliberately excludes the contact's name and title; only company-level attributes are evaluated.

Guardrails built into the product

Limitations you should know about

Account-fit estimates are probabilistic, not guarantees — treat them as one input among several, particularly for high-stakes or high-value accounts. The rationale text is generated, and generated text can occasionally be wrong or miss context a human rep would catch. Nothing here replaces your own judgment about a specific account.

Questions

hello@crmscoring.com