Underwriting & Claims // AEGIS

Price the
risk. Show
the working.

RiskPredict is an underwriting desk for digital assets. Score an asset, quote a premium, bind the policy, and run the claim end to end — on a model that returns every factor it used, so the same asset scores the same way twice and a disputed price has a line item to argue with.

Start free — 3 assessments Create an account

Paid plans open shortly — the free tier is live now

The workflow

Quote. Bind. Claim.

Four screens and six endpoints, each of which does exactly one thing you can point at.

Assessment

Score any of six asset classes — revenue stream, contract, brand, digital product, creative asset, IP claim — against a coverage amount. You get a 1–100 score, a rate band, a premium, and the itemised factors that produced them.

Binding

A quote is a stored, expiring object. Issuing a policy binds one quote by id and takes coverage, score and premium from that record — a policy cannot be written at a price the desk never quoted, and a quote can only ever bind once.

Claims

File against a live policy, get an automatic fraud score and a proposed 90/10 settlement split, then approve, settle or deny. Expired and revoked coverage refuses claims outright rather than accepting them and sorting it out later.

Forensics

Every claim carries an evidence locker and an append-only timeline. Fraud indicators are named, not scored in the dark: thin description, no evidence attached, requested over 90% of coverage, elevated automatic score.

Risk telemetry SOVEREIGN

Push signals — threat, breach indicator, fraud signal, market shift, claim surge — from your own monitoring or from other Graylight nodes, and have them move a bound policy's score while it is live. Revoked and expired policies are never re-scored.

Exposure ledger

Coverage written, premium booked, and settlements paid — with denied and still-open claims reported separately instead of folded into the totals. A refusal is never counted as revenue.

The model

Every score shows its factors

A quote returns the arithmetic, not a verdict. Six factors, each bounded, each derived from your own book.

POST /v1/aegis-ops/quote → 200
"risk_score": 58, "risk_status": "LOW_RISK",
"risk_rate": 0.05, "premium_amount": 2500.00,
"factors": [
  { "factor": "asset_class",     "points": 25,  "reason": "Base rate for asset class \"brand\"" },
  { "factor": "coverage_band",   "points": 10,  "reason": "Coverage at or above $50,000" },
  { "factor": "claims_history",  "points": 9,   "reason": "3 non-denied claims on this asset class in the last 12 months" },
  { "factor": "live_telemetry",  "points": 9,   "reason": "Net fleet risk telemetry over 30 days (+9)" },
  { "factor": "concentration",   "points": 5,   "reason": "This asset would be 62% of the active book" }
]
// Same asset, same book, same score. Every time.
Straight answer

What this is, and is not

There is no AI in the price. The score is a documented, deterministic model, and that is the point: an underwriting price you cannot re-derive is one you cannot defend in a coverage dispute. Reproducible beats mysterious.

RiskPredict is software, not an insurer. It is the desk you run your own book on — quoting, binding, claims handling and the exposure ledger. It does not carry risk, hold capital, or move money to a claimant; settlement figures are yours to act on.

It is not identity or transaction fraud scoring. Earlier versions of this page said it was. It scores assets you are covering, not requests hitting your API.

Pricing

Pay per assessment, or run the desk

The assessment is the metered unit. Policies, claims, evidence and the exposure ledger are never metered — your underwriting record is yours on every plan.

Free

$0/ month
  • 3 assessments per month
  • Unlimited policies and claims
  • Evidence locker and claim timeline
  • Exposure ledger
Create an account

Pay as you go

$15/ assessment
  • One assessment credit
  • Never expires
  • Spent only after your free three
  • No commitment
Create an account

Sovereign

$499/ month
  • Unlimited assessments
  • Risk telemetry ingest
  • Live policy re-scoring from signals
  • Portfolio exposure reporting
Start free

Billing is handled by Chancellor, Graylight's billing orchestrator. RiskPredict holds no card details and no payment credentials of any kind. Cancelling drops you to the free tier's assessment limit and switches telemetry off — it never deletes a policy, a claim or a piece of evidence.

#Omnichannel Signal Bundle

ONE COMMAND.
TOTAL REACH.

Gmail, Twilio, and RingCentral are legacy fossils. Connect the **Sovereign Signal Core** and own every communication channel in your fleet.

VOCAL
ECHO
BURST
WIRE
ACTIVATE BUNDLE — $29/mo

EXPAND THE RIG.

Tactical Synergies // Inter-Node Connectivity

View Full Manifest

VOCAL

#SmartSwitchboard

24/7 AI Receptionist. Integrate this instrument to harden your operational sovereignty.

INITIALIZE CONNECTION

KINOS

#ActiveReferee

Conflict Resolution. Integrate this instrument to harden your operational sovereignty.

INITIALIZE CONNECTION

A-OS

#NeuralCore

Business Automation. Integrate this instrument to harden your operational sovereignty.

INITIALIZE CONNECTION