AI Disclosure

1. Where automation is used

Genvernium uses automated processing for website collection, structured SEO and GEO checks, evidence organisation, findings, recommendations, and selected content or assistant workflows. Some checks are deterministic; other outputs may be generated or assisted by models when the relevant capability is enabled.

External SEO, GEO, social, or model-provider integrations are not treated as enabled merely because a feature is visible. A capability remains unavailable until it is configured and verified in the service.

2. Inputs and outputs

Inputs may include customer-authorised public website content, URLs, workspace configuration, evidence, findings, customer prompts, and review instructions. Outputs may include summaries, check results, proposed fixes, content drafts, scores based on recorded checks, and workflow suggestions.

Automated outputs can be incomplete, stale, incorrect, or biased. Customers must inspect the linked evidence, assess context, and correct or defer a finding before relying on it. Scores and recommendations do not guarantee rankings, citations, traffic, revenue, or any other outcome.

3. Human review and boundaries

Genvernium records evidence references, versions, statuses, review events, and approvals where the workflow supports them. A generated draft or recommendation does not publish a change, send outreach, make a purchase, or create another external effect. Customer approval is required before an external action is allowed.

The service is not intended for employment, credit, housing, insurance, health, law-enforcement, or other high-impact decisions about people. Do not use it to infer sensitive traits or make unlawful automated decisions.

4. Providers, retention, and training

When a third-party model or analysis provider is enabled, the applicable provider, purpose, data access, processing location, retention, training setting, and contractual controls must be assessed before activation. Unavailable providers must remain visibly unavailable, and the service must not represent placeholder values as verified measurements.

Genvernium does not use confidential customer workspace content for unrelated model training by default. Any material change to this position requires a revised notice, appropriate control, and where required a separate explicit choice.

5. Reporting an unsafe or incorrect result

To report an unsafe, inaccurate, stale, or inappropriate result, contact support@genvernium.com and include the workspace, record or evidence reference, and a description of the issue without sending secrets. We may preserve the minimum information needed to investigate, correct, or prevent recurrence.