Skip to content
Technology

Insurance,rebuilt arounddata.

1B combines underwriting, data, monitoring, technology and insurance into a single system. The output is not a faster quote, it is a fundamentally better-informed one.
L1
Signal ingestion
L2
Risk representation
L3
Underwriting decision
L4 / L5
Live programme and event feedback
Architecture

From signalto decisionand back.

Five layers, and one return path. The return path is what separates a risk platform from a risk database.
  1. L1Signal

    Ingestion

    Technical, operational and contextual inputs are collected from the systems that carry the risk: on-chain state, telemetry, control attestations, dependency graphs, threat context.

    • On-chain state
    • Fleet telemetry
    • Control attestations
    • Dependency graph
    • Threat context
  2. L2Model

    Risk representation

    Raw signal is resolved into a structured representation of exposure: what could fail, how severely, how correlated it is with everything else in the portfolio.

    • Exposure modelling
    • Severity distribution
    • Correlation mapping
    • Scenario construction
  3. L3Underwrite

    Decision

    Underwriters work on the representation, not on the raw data. Judgement stays human; the evidence base is machine-maintained and always current.

    • Appetite logic
    • Pricing view
    • Terms and conditions
    • Referral thresholds
  4. L4Operate

    Live programme

    The bound risk stays instrumented. Deviation from the underwritten state produces an action rather than a surprise at renewal.

    • Deviation detection
    • Remediation workflow
    • Portfolio aggregation
    • Re-rating
  5. L5Respond

    Event

    Incident data returns to the model. Every event, insured or not, improves the representation for the entire portfolio.

    • Incident capture
    • Loss attribution
    • Model feedback
    • Portfolio learning

Evidence returns to the model, the portfolio learns from every event

1B Risk OS

Your datashould negotiatefor you.

A permissioned intelligence layer between what is verifiably true about a risk and what an insurer can act on. It turns evidence the customer controls into a view of the exposure an underwriter can read, and it keeps the record of how every conclusion was reached.
Evidence rises through the stack; what is learned returns to the bottom of it. That return path is the difference between a risk platform and a risk database.

Architecture

From reality to insured risk.

The data layer and the insurance layer are deliberately separate. One establishes what is true and who may use it. The other decides what that means for an exposure, and carries it to a market that still has to approve it.

  1. RealityPeople, businesses, assets and machines. Source systems, behaviour, ownership, telemetry and records.
  2. TrustIdentity, provenance, lifecycle state, verification. Establishing what is actually true before anyone acts on it.
  3. PermissionPurpose limitation, selective disclosure, authorisation, revocation. The customer decides what leaves and what it is used for.
  4. Risk TwinThe exposure as insurance sees it: loss drivers, positive signals, gaps, and what a carrier will require.
  5. Underwriting intelligenceEligibility, structure, coverage requirements, and the questions that have to be answered before a risk can be placed.
  6. AgentPrepare, compare, explain, monitor. It does the assembly work. It does not make the decision.
  7. MarketBroker, MGA, carrier, reinsurer. Appetite, capacity, approved wording and human underwriting.
  8. ProtectionCover bound by a licensed carrier, subject to its approvals and to the terms of the policy.

Two questions, two layers

What is true?

The trust layer connects and verifies the economic context. It establishes provenance and permission, and it deliberately does not make the insurance decision.

  • Identity and ownership
  • Asset record
  • Current state and telemetry
  • Lifecycle evidence
  • Who may use what, and why

What does it mean for insurance?

1B turns authorised evidence into a view of the exposure an insurer can read, and orchestrates the path from submission to capacity.

  • Insurability
  • Risk structure
  • Coverage requirements
  • Market fit
  • Decision provenance
Private Risk Twin

Data is notthe product. Thereading of it is.

A digital twin describes the object. A risk twin says what the object means for an insurer — and keeps, for every material conclusion, a record of what kind of statement it is.
  1. Verified fact

    Signed, current, and traceable to the credential it came from.

  2. Derived attribute

    A property computed from records that themselves stay private.

  3. Model inference

    A pattern read across a fleet or a portfolio. Labelled as inference, never as fact.

  4. Carrier rule

    A boundary the market sets. Not negotiable by the system that reads it.

  5. Human decision

    An underwriter's judgment, with the conditions attached to it.

The labels are the point. A system that cannot say which of these a conclusion is cannot be audited, and an underwriter is right not to trust it.
Data and permissions

The minimumthat answersthe question.

Every step is a reduction. What reaches an insurer is the least that answers what the insurer actually asked.
  1. Raw dataStays at source wherever it can.
  2. VerificationThe trust layer confirms only the permitted facts.
  3. Verified attributeThe minimum necessary evidence, not the record behind it.
  4. Insurance signalWhat that evidence means for the exposure.
  5. DisclosureAuthorised information only, for the request it was authorised for.
Sensitive personal information is processed only where that is lawful and properly authorised. Privacy-enhancing technology does not make otherwise prohibited underwriting lawful.
Why it compounds

Better evidenceimproves morethan the price.

The claim is not that software makes insurance cheaper. Better risk selection, less information asymmetry, less manual servicing and stronger portfolio intelligence each improve a different part of how insurance works.
  1. Embedded distributionLess friction to reach
  2. More insured riskBroader observed exposure
  3. More verified evidenceLess uncertainty
  4. Sharper segmentationRisks told apart
  5. Better underwritingMore precise eligibility
  6. Loss intelligenceClaims and near misses
  7. Capacity relationshipsPortfolios that are understood
  8. AuthorityBoundaries set per product
  9. Products and termsResidual risk better structured
  10. More distributionAnd round again
Illustrative architecture. Nothing here is a quotation, and no cover exists until a licensed carrier has approved it and a policy has been issued.
The risk object

An exposure, resolved into something priceable.

Raw signal is not underwriting input. It becomes input once it is resolved into a structured representation: what can fail, how severely, how correlated it is with the rest of the portfolio, and whether the state today still matches the state that was underwritten.
The schema below is illustrative of the model. It contains no client data.
Risk objectMAINTAINED
  • architecture.classVERIFIED
  • controls.baselineVERIFIED
  • dependency.graphMONITORED
  • autonomy.scopeMONITORED
  • deviation.signalDRIFT
  • aggregation.clusterMONITORED
  • incident.historyVERIFIED
  • remediation.queueOPEN
Deviation from the underwritten state produces an action
Signals by risk class

Different risksproduce differentevidence.

There is no universal telemetry for emerging risk. Each class is instrumented on its own terms.
Digital Assets
  • Custody architecture class
  • Quorum and signer topology
  • Cold / warm / hot ratio
  • Attestation cadence
  • Dependency concentration
  • Chain and asset exposure mix
  • Incident and near-miss history
MiCAR / CASP
  • Authorisation status and scope
  • Client asset segregation model
  • ICT and third-party register
  • Continuity and exit planning
  • Governance and fit-and-proper posture
  • Regulatory change exposure
AI Risk
  • Autonomy level in production
  • Evaluation and red-team cadence
  • Human oversight coverage
  • Tool and spend authority scope
  • Model and provider change rate
  • Incident and near-miss telemetry
  • Data provenance posture
Robotics
  • Operational design domain
  • Fleet hours and utilisation
  • Intervention and disengagement rate
  • Software version distribution
  • Proximity to human operators
  • Certification and standards posture
  • Field incident classification
Tokenized Assets
  • On-chain / off-chain reconciliation
  • Issuance control topology
  • Underlying custody arrangement
  • Oracle and pricing dependency
  • Redemption and transfer mechanics
  • Jurisdiction of the legal claim
Cyber & Digital Risk
  • External attack surface
  • Identity and access posture
  • Vendor and dependency concentration
  • Backup and restoration capability
  • Credential and exposure intelligence
  • Sector threat context
Principles

What the systemdoes, anddoes not do.

Judgement stays human

Models structure the evidence. They do not decide appetite, terms or price. An underwriter does, with a better picture than the process previously allowed.

Evidence over declaration

A control that cannot be observed is treated as a control that may not exist. Verification precedes capacity.

Current, not annual

A risk picture that is refreshed once a year is wrong for most of the year. The state is maintained continuously.

Portfolio memory

Every incident, insured or not, improves the representation for everything else in the book. The system compounds.

Data minimisation

We ingest the signals that change an underwriting decision, and no more. Scope is agreed in writing before anything is connected.

Client control

Counterparties see the same picture we do. Nothing is assessed against evidence a client cannot inspect.

Technology

Underwriting isan informationproblem. We treatit as one.

Talk to 1BExplore solutions