Skip to main content

Legal

AI System Card

Operational transparency for AI features used in SkepticFlow demos, lead intake, and product workflows.

Last updated: 2026-05-27

Systems and models

SkepticFlow may use OpenAI, Anthropic, or Azure OpenAI models for lead intake, chat assistance, summarization, routing, and draft generation.

The public chat experience identifies itself as an AI-assisted flow and offers human follow-up through contact channels.

Data flow

User messages and qualification answers can be sent to model providers and stored with session metadata to support follow-up and auditability.

Sensitive production, health, legal, payment-card, or special-category data should not be submitted unless a signed agreement explicitly covers it.

Refusals and limits

The system should refuse or de-escalate unsafe requests, requests outside business qualification, and attempts to extract credentials or secrets.

The system is not intended for autonomous legal, medical, financial, employment, or high-risk AI decisions.

Human oversight

Human review is used before project commitments, pricing, legal positions, or production implementation decisions.

Oversight contact: hello@skepticflow.com.

Reference links