Customer data boundary
Customer brands, audits, prompts, AI files, reports, billing actions, and data controls require a signed-in Google/Keycloak session and are scoped to that account.
Trust
Foundiqo trust, safety, and data posture
Foundiqo keeps SEO + GEO work grounded in crawler safety, tenant scoping, human review, data controls, and no-guarantee language.
Crawler safety
Public crawl uses SSRF-aware validation, private-network rejection, redirect checks, byte limits, timeouts, and responsible user-agent behavior. Deep crawl requires domain verification.
Generated output review
Brand Brain, Proof Graph, Offer Graph, schema, llms.txt, well-known JSON, reports, and MCP/action scaffolds stay draft or approval-gated until reviewed by a human.
Prompt and AI provider policy
Prompt outputs are evidence snapshots. Foundiqo records provider/source, answer text, citations, competitors, confidence, and limitations without claiming permanence.
Billing and payment posture
Foundiqo routes paid plan changes through provider-backed checkout handoff. It does not store card numbers directly in Foundiqo app records.
Secrets and runtime
Runtime credentials are expected from encrypted runtime secret flow. API payloads and public files must not expose provider keys or secret-like values.
Retention Reference
Page snapshots
Default 90 days unless customer policy or enterprise retention overrides it.
Prompt runs
Default 180 days for evidence review, trend checks, and report traceability.
Reports and generated assets
Stored until deleted, subscription ends, or workspace retention policy removes them.
Operational logs
30 to 90 days depending on infrastructure log class and incident needs.
Public Trust Statements
- Foundiqo does not guarantee AI rankings, citations, mentions, recommendations, conversions, or revenue.
- Foundiqo does not publish generated files without customer approval.
- Deep crawling, scheduled monitoring, and publishing actions require ownership verification.
- Competitor analysis must use public pages and respectful low-volume access.
- AI answer engines change frequently, so monitoring is probabilistic and evidence must be timestamped.