diff --git a/docs/architecture/architectural-principles.md b/docs/architecture/architectural-principles.md
new file mode 100644
index 0000000..219954e
--- /dev/null
+++ b/docs/architecture/architectural-principles.md
@@ -0,0 +1,53 @@
+# Architectural Principles
+
+---
+
+## Principle 1: Mission Scope And Contract Scope Are Separate
+
+The company may identify any digital issue that affects customer acquisition, visibility, trust, or conversion.
+
+The recurring service only includes monitoring, diagnosis, prioritization, reporting, and approved low-risk optimizations.
+
+Any work outside the defined recurring service areas is treated as a separate project unless explicitly included in the client's package.
+
+---
+
+## Principle 2: AI Visibility Is Core, But AI Placement Is Not Guaranteed
+
+The company may monitor and improve the digital signals that help AI-assisted discovery systems understand and represent a business.
+
+The company does not guarantee that any AI system will recommend, rank, cite, or surface a client in response to any specific query.
+
+---
+
+## Principle 3: Reports Are Receipts, Not The Product
+
+The product is continuous digital oversight and optimization.
+
+Reports exist to document what was monitored, what was found, what was improved, what remains at risk, and what is recommended next.
+
+---
+
+## Principle 4: The Company Improves Digital Pathways, Not Every Marketing Function
+
+The company focuses on the digital pathways that turn customer interest into customer action.
+
+The company does not become a general-purpose marketing agency by default.
+
+---
+
+## Principle 5: Verified Evidence And Indicative Evidence Must Be Separated
+
+The company distinguishes between verified defects and indicative risks.
+
+Verified defects may be presented as confirmed issues when supported by concrete evidence.
+
+Indicative risks must be presented as directional evidence, not proven customer loss.
+
+This distinction protects client trust, prevents overclaiming, and keeps reports credible.
+
+Agents may propose evidence classifications, but humans approve final evidence tiers before client-facing use during Version 1.
+
+Ambiguous findings default to Indicative Evidence unless upgraded by human review.
+
+Rejected Verified Evidence classifications default to Indicative Evidence unless the reviewer discards the finding or requests additional evidence.
\ No newline at end of file