About

An observatory for checkable AI answers

AnnuityRates.AI exists to publish reproducible research about what AI assistants say—not to sell annuities, recommend products, or collect prospects.

Mission and boundaries

The observatory tests a fixed set of consumer-realistic questions and evaluates answers against disclosed sources. It publishes no annuity rates or rate tables, no carrier or product pages, no product recommendations, and no quote, matching, phone, or lead form.

Accountability

Accountable operator pending verification.

The named operator is a Galen approval gate. Until an approved primary-source record supports the operator identity, this site renders no biography, credential, byline, or reviewer chip.

Sources and generation

Question rationales and grading plans are authored before collection. Expected claims are sourced to IRS publications, state guaranty associations and regulators, and verified sister-property explainers where they add editorial context. Benchmark answers will be collected through approved APIs only after spend approval. Scores will be computed from the frozen rubric rather than asserted by editors.

Independent review

Registry rows begin as “pending independent review.” A named reviewer can appear only after a signed memo, written naming consent, and durable direct primary-source verification are attached. One reviewer is permitted per page.

Cite us and press

No benchmark result is available to cite yet. When run #1 is published, its report package will provide a stable title, version, publication date, methodology link, transcript archive, and dataset citation. Press inquiries and contact details are withheld until an accountable operator and channel are approved.

Analytics status

The site contains a dedicated-project, production-host-gated PostHog integration with masked inputs, console capture disabled, and surveys disabled. It remains off when the site-specific public ingestion token is absent; project provisioning and credentials are owed.