Procurement · answered

Data, IP and retention. In one place.

What Amoris reads. What it writes back. What it never asks for. Who owns what at the end. If your legal team has a question that is not here, they can send it and it will be answered.

Why trust matters here

Confidence cannot exceed evidence.

How Amoris prevents AI hallucinations in GTM decisions.

01

Every claim gets a label.

FACT (verified), INFERENCE (logical from evidence), HYPOTHESIS (plausible), UNKNOWN (not enough evidence).

02

Confidence calibrated to evidence.

90% only when multiple sources agree. 10% when the claim is speculative. No overreach, no hallucinations.

03

Reasoning is auditable.

Every claim carries its source and rationale. Your legal team can trace any decision back to what it was built on.

The four procurement questions

What does Amoris read from us?

Company documents, prior GTM output, sales collateral, CRM export, and optionally a CSV of dormant leads. Staged access — nothing is asked for that a phase does not need.

Where does the data live?

On your VPC when the engagement supports it. In a scoped tenant otherwise. Every table and every graph node is tenant-scoped; retrieval refuses to run without a client scope.

What will Amoris never request?

Customer email content, individual contact records outside the scoped list, payment or banking data, or credentials to any consequential system. Human approves the sends; Amoris does not need the account.

Who owns what at the end?

You own the map, the data, and the outputs. Amoris retains a perpetual irrevocable royalty-free licence to the methods, procedures and ontology — the compounding memory is the moat, not your data.

The evidence framework

Truth before traction

Confidence cannot exceed evidence. No overreach. No hallucinations.

Truth before traction: evidence classification and confidence gate

PDF · Free

Free download

The AI Evidence Framework

How Amoris prevents AI hallucinations by grading every claim on evidence before it drives a GTM decision.

  • The 4-tier evidence classification (FACT / INFERENCE / HYPOTHESIS / UNKNOWN)
  • The confidence calibration table, 10–90%
  • A worked example applying the framework to a real GTM claim

We only email about Amoris. No newsletter. Unsubscribe with one reply.

Retention policy and DPA

Shipped before any third-party PII enters the system.

A retention policy and a mutual DPA are prerequisites of accepting third-party lead PII. Until they land, the optional CSV upload on /diagnostic is disabled and the scoping call runs on your description of the funnel.