Policy

AI Governance & Transparency

How AEOfix uses artificial intelligence — what it does, what data it touches, who reviews it, and what rights you have.

Last updated March 18, 2026.

1. How AI Is Used at AEOfix

AEOfix uses Google Gemini 2.0 Flash via API to generate text in the following features:

  • AEO Audit Reports — Schema and content recommendations produced by a three-agent pipeline (schema analyst + content architect + verifier). This is an operator-level tool used by AEOfix to audit its own site's performance for each AI engine. It is not a self-service feature for customer sites.
  • Weekly Bot Intelligence Digest — Narrative summary of your site's bot visit trends, delivered by email each Monday
  • Monthly Intelligence Brief — Broader AI crawler behavior report delivered on the 1st of each month
  • Page Optimization Suggestions — Top 5 AI-crawled pages with improvement recommendations, delivered with the weekly digest

What Gemini does not do: Gemini does not control any feature of your site, your account, or any decision affecting your business. It writes suggestion text. You decide what to do with it.

2. What Data Is Sent to AI

When AI features run, the following data is transmitted to the Gemini API:

  • Aggregated bot visit statistics — counts, page paths, timestamps, user-agent strings, geographic region codes — for your registered site
  • Publicly accessible HTML content from your tracked pages (fetched by standard HTTP — identical to what any browser visitor sees)
  • Structured prompt instructions authored by AEOfix

Never sent to Gemini:

  • Your name, email address, or any information that identifies you personally
  • Payment information, billing records, or subscription details
  • Passwords, authentication tokens, or session data
  • Any content from your site that is not publicly accessible via HTTP

Data sent to Gemini is processed transiently for report generation. Google's API terms prohibit use of API inputs to train their models without operator consent — AEOfix has not granted that consent.

3. Human Oversight

Every AI output AEOfix generates goes to a human reader — you — before any action is taken. AEOfix has no automated pipeline that reads Gemini output and does something to your site or account without your review and decision.

The AEO Audit feature uses a three-stage pipeline specifically designed to reduce AI error before it reaches you:

  1. Schema Analyst — identifies schema opportunities on the page
  2. Content Architect — evaluates content structure and AEO alignment
  3. Verifier — cross-checks both outputs against the actual live page and produces an explicit confidence score: HIGH, MEDIUM, or LOW

A LOW or MEDIUM confidence score means the recommendation should be treated as a starting point for your own judgment, not a directive. Review all AI outputs before acting on them.

4. What This AI System Is Not

The following are explicit statements about AEOfix's AI use, not marketing language:

  • Not a high-risk AI system under Texas TRAIGA (SB 2060), the EU AI Act (Regulation 2024/1689), or Colorado SB 24-205. AEOfix does not make or materially influence consequential decisions about employment, housing, credit, insurance, healthcare, education, or legal rights.
  • Not autonomous. Gemini has no ability to take any action on your behalf. It generates text. A human reads that text and decides.
  • Not trained on your data. Your site data is processed transiently and is not used to train the Gemini model or any AEOfix model.
  • Not personally profiling you. No personally identifying information enters any AI prompt.
  • Not making decisions about your account. Access, billing, and account status are governed by rule-based logic — not AI outputs.

5. Framework Alignment

AEOfix maintains internal governance documentation aligned with the NIST AI Risk Management Framework 1.0 across all four functions. Here is what each function means in practice for our system:

Govern

Written policy, single accountable owner (William J Bouch), documented AI purpose and scope, annual review schedule. No AI feature is added without updating the internal governance document.

Map

Formally characterized as low-risk text generation for marketing analytics. Risk level assessment documented: domain is non-consequential, customer retains all decision authority, no protected categories involved.

Measure

Verifier agent produces explicit confidence scores (HIGH/MEDIUM/LOW) displayed to you. Owner performs quarterly manual spot-checks of audit recommendation accuracy. Customer complaints are logged and inform reviews.

Manage

Customer opt-out available for digest narrative and page suggestions. AI output disputes acknowledged within 48 hours, resolved within 10 business days. Provider change requires 14-day advance customer notice per Terms §16.6.

Texas TRAIGA: AEOfix operates as a deployer (not developer) of a third-party AI model (Google Gemini), outside all high-risk categories defined by SB 2060. The affirmative defense is maintained through documented, substantially equivalent NIST AI RMF 1.0 compliance adopted prior to any alleged violation. Google LLC is independently responsible for developer-level obligations under applicable AI law.

6. Your Controls

Opt Out of AI Features

You may disable AI-generated narratives and recommendations at any time via Notification Preferences in your dashboard settings. Bot visit data will continue to be collected and displayed in tabular form without AI processing.

Dispute an AI Output

If you believe an AI-generated recommendation was inaccurate, misleading, or harmful:

  • Email AEOfix.com@gmail.com with subject line: AI Output Dispute
  • We will acknowledge within 48 hours and resolve within 10 business days
  • If the output was demonstrably inaccurate or caused harm, we will remove it from your view and issue a service credit or refund as appropriate

Further Reading