Frequently asked questions

What does MortarIQ do?

MortarIQ connects to your data warehouse with read-only access, reads schema metadata, and scores your data's readiness for AI workloads across six factors. You get a readiness verdict, a requirement-level breakdown, a PII exposure inventory, prioritized recommendations, and remediation SQL you can hand to your data team.

Is it safe to connect my warehouse?

The connection is read-only and metadata-only. We query system catalogs (INFORMATION_SCHEMA and platform system views) and never run SELECT against your tables, so we never see your rows, values, or query results. Every query we run is published verbatim at /security/queries, generated from the connector source code so the page cannot drift from what actually executes. Each query is also tagged with a per-scan id so your database administrator can verify it in your query history.

What permissions do I need?

What do you send to your AI provider?

Assessment results: scores, requirement outcomes, and the metadata-derived diagnostics they cite, which can include database, schema, table, and column names. That specificity is what makes recommendations actionable, for example naming the exact column to mask. We never send your data values and never your credentials. Our AI provider (Anthropic) does not train on API inputs.

Do you train AI models on my data?

No. Your metadata, assessment results, and credentials are never used to train models, generate cross-customer benchmarks, or improve the product for other customers.

Why do some checks say "not assessable"?

Because we refuse to fake a number. Some requirements need actual data values (data completeness, uniqueness), some live in a governance catalog rather than the warehouse, and some are not exposed by a platform's catalog. Where we cannot measure something we say so, the score renormalizes over what was measured, and the report tells you how many requirements were measured. A high score on measured requirements never silently claims coverage of unmeasured ones; the maturity label is capped when a heavily weighted factor could not be measured.

Which compliance frameworks do you map to?

Seven selectable lenses: EU AI Act Article 10, NIST AI RMF, ISO/IEC 42001, ISO/IEC 5259 (ML data quality), GDPR, SOC 2, and HIPAA. You select the frameworks your organization tracks and every assessment maps its findings to that framework's data-governance expectations: evidence in place and gaps to close. This is readiness to produce evidence, never a certification claim.

Which platforms are supported?

Google BigQuery, Snowflake, Databricks, PostgreSQL, Amazon Redshift, and Microsoft Fabric. Snowflake and Databricks work on any cloud they run on. Managed PostgreSQL (Amazon RDS and Aurora, Cloud SQL, Azure Database for PostgreSQL, Neon, Supabase, Railway, Heroku) connects through the PostgreSQL connector.

Lakehouses are covered through their catalogs: Databricks is assessed via Unity Catalog, Microsoft Fabric works against both Warehouse and Lakehouse SQL analytics endpoints, and Iceberg or external tables registered in Snowflake and BigQuery appear in the same catalog views we already scan. Data sitting in object storage outside any catalog is not visible to a metadata-only scan.

My warehouse is not reachable from the internet.

Two options. Allowlist MortarIQ's egress IP in your firewall or security group, or run the assessment inside your own network with the CLI agent, where credentials and the connection never leave your machine. Add --html for a shareable local report (no account needed), or optionally upload only the computed results to your dashboard.

What does it cost?

There is a free plan (two assessments per month, scan profile, latest result) with no credit card required. Pro and Team add the deep workload profiles, AI recommendations, history, scheduling, and the remediation board. Pricing is on the landing page; there is no sales call.

How is this different from data observability tools?

Observability tools (Monte Carlo, Soda, Anomalo) continuously monitor data values for engineers who already run pipelines. MortarIQ answers a different question for a different audience: can we responsibly build AI on this data, asked by the data leader before the build starts. You get a go/no-go verdict per AI workload, an honest account of what could and could not be measured, compliance lens mapping, and remediation steps, from a read-only metadata scan that takes minutes.

Can I delete my data?

Yes. Delete a saved connection from Settings at any time; credentials are permanently removed, not soft-deleted. Request account deletion at support@intellibricks.app and we delete your data within 30 days. Details are in the Privacy Policy and DPA.