Audit-grade AI for
underwriting and claims.
Aegis pairs a deterministic insurance risk engine with an explainable AI layer. Every decision is grounded in your rules, logged with full rationale, and ready for state DOI examiners, internal audit, and your reinsurance partners.
- SOC 2 Type II ready
- NAIC Model #672 aligned
- HIPAA-aware data handling
- Immutable audit log
POST /decide/underwrite
Personal auto · age 19 · high-risk profile
Risk score
77/100
Confidence
72%
Suggested premium
$124.10
Model
gpt-4o-mini
Rationale
- ▸Base rate for 'auto': 0.060
- ▸Policy type 'auto' classified as elevated risk (+10)
- ▸Young applicant (age 19) increases risk (+12)
- ▸Declared risk profile: HIGH (+20)
Platform
Everything underwriting and claims teams need
to deploy AI responsibly.
Deterministic Risk Engine
Actuarial-style rules compute risk and fraud factors before any AI call. The model decides on top of facts, not free-form text.
Audit-Grade Decision Log
Every decision is persisted with inputs, factors, model, version, confidence, and full rationale. Built for state DOI and internal audit reviews.
Self-Hostable & Offline
Run on your own infrastructure, in your VPC, or fully air-gapped. No customer PII has to leave your perimeter.
Underwriting Workbench
Risk score, suggested premium, and an explainable rationale on every new policy submission across auto, home, life, and commercial lines.
Claims Triage
Fraud scoring, out-of-window detection, severity flags, and a clear APPROVE / FLAG / REJECT recommendation for every FNOL.
Model-Agnostic
Plug in OpenAI, Anthropic, Azure OpenAI, or a private on-prem model. A heuristic fallback keeps the platform deterministic when you need it to be.
Architecture
Rules first. AI second.
A deterministic risk engine computes scores and produces structured factors using your carrier's rules. That block is then passed to the language model as grounded context. The model's only job is to issue a decision and summarize the rationale, which is written to an append-only decision log you control.
rule_engine
└─ risk_score, factors[], suggested_premium
│
▼
llm(system + structured_context)
└─ { decision, confidence, reasoning[] }
│
▼
Decision table ◀── audit log, immutable
│
▼
GET /decisions ─── compliance & dashboardsCompliance & Security
Designed for the
regulated enterprise.
Aegis is built around the controls examiners and internal audit teams expect from a North American carrier, not a consumer chatbot.
Immutable audit trail
Append-only decision log with input snapshot, model id, and version on every record.
Deterministic fallback
Heuristic engine keeps the platform answering even when the LLM is unavailable.
Self-host or VPC
Deploy on your own cloud or on-prem; no customer PII is required to leave your perimeter.
Explainable rationale
Plain-English reasoning attached to every APPROVE / FLAG / REJECT decision.
Role-ready APIs
Stateless REST endpoints fit cleanly behind your existing IAM, SSO, and rate limiting.
Open source core
MIT-licensed engine you can audit line by line; no vendor black box.
Decisions your underwriters,
auditors, and counsel can defend.
Open the dashboard, submit a sample policy, and watch the underwriting agent reason about it in real time. No credit card and no API key required to evaluate.