§ Trust & compliance
Defensibility is not a feature — it is the product. Six pillars structure how Verity Council keeps AI consultation safe to rely upon.
01 · Independence
Each expert maintains a Restricted Entity List. Each firm workspace maintains its own. Every consultation runs a real-time check against both — blocking or warning before the agent answers, and logging the result for audit.
02 · Knowledge currency
Every agent is tagged with a standard version and effective-date range. When standards update, affected agents flip to Legacy Mode automatically — answers carry a Superseded Standard banner until the expert retrains.
03 · AI disclosure
Every interaction is labelled: AI-only, AI with human review, or live expert. Every memo records the agent ID, version, skepticism setting, and reviewer signature where applicable.
04 · Indemnity tiers
Tier I (AI-only) — standard disclaimer. Tier II (AI + human review) — shared liability with platform-backed E&O up to a cap. Tier III (live expert) — full professional engagement with expert-carried insurance verified at onboarding.
05 · Evidence locker
Agents anchor logic to specific paragraphs of the FASB Codification, IFRS Blue Book and applicable auditing standards — and surface those anchors in every response and memo.
06 · IP protection
Experts certify they upload only original work. An IP scanner flags suspected proprietary content (Big 4 manuals, client files, internal training material) before it can enter training data. Confidentiality, NDA and data-retention controls apply throughout.
Required disclosure
"This output is AI-assisted technical consultation support and does not replace the professional judgment of a licensed professional."
Carried on every consultation. Agents are not licensed professionals; they are trained using licensed expertise from a verified consultant.