Frequently Asked Questions
Find answers to common questions about Speculum AI, Mirror OS, and our governance infrastructure.
Buyer FAQ
For decision makers and business leaders
Mirror OS is the governance infrastructure for agentic AI. It governs decisions at boundaries — before, during, and after execution — and generates regulator-grade evidence by default.
Regulated enterprises deploying agentic AI — particularly in financial services, insurance, healthcare, and government. Organizations where AI failures create compliance violations, not just technical issues.
If your AI systems make autonomous decisions, interact with customers, or operate in regulated workflows — and you can't fully explain what they did or why — you're ready.
Observability tools monitor what happened after the fact. Mirror OS governs what's allowed to happen in real-time, and generates audit-grade evidence by default.
Before: You react to AI incidents after they happen. After: You govern decisions before execution and can explain outcomes with evidence.
No. Mirror OS is modular. Most teams start with one module for a specific workflow, establish a baseline, then expand step-by-step.
SaaS, VPC, and on-prem depending on your security and compliance requirements.
Usage-based pricing tied to decision volume, with enterprise agreements available. Most customers start with a short pilot on a high-risk workflow.
Trust & Technical FAQ
For security, compliance, and engineering teams
Speculum AI builds governance infrastructure for agentic AI — controllable, auditable, and safe to scale.
Guard enforces policy at the interaction level. Gateway manages execution routing and orchestration across providers.
Via API gateway, SDK, or proxy deployment. Works with any LLM provider and any orchestration framework.
Policy enforcement typically adds <50ms. Designed for real-time governance without killing performance.
Yes. SaaS, VPC, and on-prem options available.
Tamper-evident decision records: input/output pairs, policy evaluations, risk scores, timestamps, attribution chains, and human overrides.
SOC 2, ISO 27001, GDPR, aligned with EU AI Act and J-AISI guidelines.
Optimized for agentic AI and LLM-based systems. Supports multi-agent workflows, RAG pipelines, and autonomous decision chains.