Frequently asked questions
Common questions about scoring, security, languages, and pricing.
6 min read
Is my data used to train AI models?
No. We hold zero data retention contracts with both Anthropic and Cohere. Your documents, scores, and recommendations are never used to train any model. Every model call is API-only with no fine-tuning opt-in.
Where is my data stored?
By default in me-central-1 (UAE). EU region available on the Continuous tier. You choose at workspace creation and we enforce it via Supabase region binding — no cross-region replication.
Can I dispute a score?
Yes. Every AI score has a Dispute this score button that opens a context field. Add new evidence + reasoning, click rescore, and the question is re-evaluated against the updated evidence set. Disputes are logged in the audit trail.
How long does a full assessment take?
With 30+ documents uploaded, bulk-scoring all 216 questions takes 15–30 minutes of model runtime. Preparation (gathering documents, inviting contributors) usually takes 2–5 working days. Compare with 8–12 weeks for a consultant-led assessment.
How is this different from consultants?
A consultant writes a score on a slide based on a one-hour interview. We give you a score linked to a specific quoted passage from your own document, auditable + disputable. Both can be useful — but only one is defensible to your QA team without re-running the analysis.
Do you support SAML SSO?
Currently we offer Google + Microsoft OAuth out of the box. SAML (Okta, Azure AD, Ping) is on the roadmap for the Continuous + White-Label tiers. Contact us if you need it sooner.
What about ISO 27001 / SOC 2?
ISO 27001 is in progress (we're going through our own certification cycle). SOC 2 Type II is on the 2027 roadmap. Until then we provide standard DPA, sub-processor list, and a security questionnaire response on request.
Can I export my data?
Yes — at any time. Owner role can export the full assessment (answers, scores, evidence library metadata, recommendations) as JSON / CSV / PDF. Tenant deletion is a hard delete: 7-day grace, then evidence, chunks, embeddings, scores, and audit logs are purged from primary + backups within 30 days.