Secoda AI is built on a foundation of enterprise-grade security including SOC 2 compliance, encryption in transit, and comprehensive audit trails to ensure your data remains protected and compliant.

No, your data is never used to train AI models. This is one of Secoda's core security commitments, backed by multiple layers of protection. By default, only metadata is sent to Secoda AI. This includes resource names, descriptions, documentation, ownership information and tags, popularity metrics and usage statistics, and schema information and relationships.

Yes. Administrators can set up governance filters to specify exactly which resources AI can access through inclusive or exclusive filter rules. You can also disable specific AI tools like SQL execution on a workspace level.

Secoda AI strictly adheres to team-based data access permissions. If a user asks about resources from a team they're not part of, AI will not return any results. This ensures access remains secure and relevant to the user's team assignments.

By default, Secoda AI only accesses metadata from your workspace, including resource names, descriptions, owners, tags, definitions, and popularity information.

Secoda AI can be configured to respect data classification tags (PII, confidential, sensitive), exclude sensitive resources through governance filters, and ensure columns marked as PII are not displayed in data previews.

Yes. Each persona can be configured with specific team access controls, ensuring only authorized users can access role-specific AI assistants.

No. Each persona operates within its configured boundaries and cannot access resources outside its assigned scope. Persona interactions are isolated and follow the same governance rules as regular AI interactions.

Advanced Memory allows Secoda AI to learn from past successful interactions within your workspace to improve response quality. This learning remains within workspace boundaries, analyzes interaction patterns to identify successful strategies, builds workspace-specific knowledge tailored to your data pattern, and can be managed, controlled, or disabled by administrators.

Unlike some AI services that may use interactions to improve their models, Secoda's approach ensures your conversations remain private and are not used for training, your data patterns don't influence the base AI models, your business context stays within your organization, and your competitive information remains confidential.

Some data analysis use cases for Secoda AI include:

Faster analysis with natural language : Ask plain-English questions like “What’s the churn rate for Q4?” and get guided to the right tables, queries, or dashboards.

Smart query assistance: Speeds up query building, especially when navigating complex schemas or joining unfamiliar tables.

Faster time to value: Analysts can respond to ad-hoc stakeholder questions faster, with fewer clarifying back-and-forths. If stakeholders have access, they can even self-serve on common data questions using AI.