Public information first
An initial AI Search Snapshot uses public business information and does not require account credentials. Public facts can still be wrong, so conflicts remain labeled until an authoritative source resolves them.
Published safeguards
AI and client data policy
The practical rules Sequoia GEO uses when AI tools, public research, client accounts, and customer information meet in the same engagement.
Version 1.0. Published August 30, 2026.
The short version
Sequoia minimizes the data shared with an AI provider, keeps credentials out of prompts, avoids personal customer data in public AI tools, and requires human review before a recommendation becomes a client-facing claim or live change.
Credentials stay out of model prompts
Passwords, API keys, private keys, access tokens, and recovery codes are not copied into prompts. Access is granted through the platform's normal permission system when an engagement requires it.
Customer data is not public prompt material
Personal lead and customer details are not placed into public AI chat tools. When analysis requires customer-level records, Sequoia limits the data to the approved purpose and uses de-identified summaries when that is sufficient.
Human review remains required
AI output is treated as a draft or research aid. A person reviews factual claims, client-facing content, recommendations, and proposed live changes before they are published or implemented.
Named examples require approval
A client or prospect is named in a case study, assessment example, testimonial, or public result only with permission. The published page states the evidence period and does not turn an observation into a guaranteed outcome. If that permission is withdrawn, Sequoia removes the named example from its current public pages and public knowledge catalog.
Clients retain ownership
Client accounts and client-provided data remain the client's. Sequoia uses approved access for the agreed scope and does not claim ownership of the client's advertising, analytics, CRM, website, or business records.
Third-party AI systems have their own terms
Retention, training, residency, and deletion controls differ by provider, product, and account configuration. Sequoia does not promise zero retention or a deletion timeline unless the selected provider and configuration support that promise and it has been verified.
If an engagement needs sensitive account or customer data, the required system, fields, access method, and purpose are agreed before that data is used. A public-surface review remains available when account access is not appropriate.