Straight-through processing (STP) in insurance is the handling of a submission, from intake to a bound decision, with no manual touch by a human underwriter. In Seguros e Danos (P&C, property and casualty), an STP case is one where a quotation request enters the pipeline and is read, scored, priced, and then quoted, declined, or issued automatically, because the risk is low complexity and sits clearly inside the insurer's risk appetite and underwriting manual (manual de subscrição). The defining feature is not speed alone. It is the absence of human intervention on the cases the insurer has already authorized rules and models to decide.
STP does not mean approving everything automatically. A well-designed flow auto-declines and escalates with the same legitimacy with which it quotes. The decision to send a complex or out-of-appetite risk to a human is itself part of the straight-through logic. The pipeline routes in-appetite, low-complexity risk to automatic quotation or issuance, escalates out-of-appetite or high-complexity risk to an underwriter, and sends fraud signals or data gaps to investigation or enrichment. The underwriter is freed for the cases that genuinely need judgment.
It is worth separating two close terms. A touchless underwriting pipeline usually describes the end-to-end capability and the goal state, where the digital pipeline can handle a submission with no human touch. The STP ratio is the operational measure of that capability in production. It names the share of real submissions that actually completed without a touch. Put another way, the touchless pipeline is the design and the STP ratio is the scoreboard. A line of business can be built to run touchless and still show a modest STP ratio, because many real submissions are still referred to a human.
How STP works across the underwriting flow
The STP pipeline runs through linked stages, each with a full audit trail. First comes multichannel intake: submissions captured from broker (corretor) email, PDF, portal, and messaging, normalized into a single structured pipeline, in the format the insurer already uses. Next, intelligent document reading extracts risk data from the proposal, prior policies, and schedules, removing the manual re-keying that consumes underwriting hours before any analysis begins.
Then enrichment and scoring complete missing data from internal and external sources, cross-referencing information such as company registration (CNPJ), broker history, exposure, and credit, and score the risk against the appetite. The fourth stage is the risk and fraud engine, a multi-factor Machine Learning model calibrated to the insurer's risk appetite and underwriting manual, which assesses risk quality and flags fraud signals. Dynamic pricing follows, with the premium (prêmio) calculated against the scored risk and an instant output.
Finally, decision and prioritization routes each case. The pipeline quotes, declines automatically, or escalates to a human underwriter, always with an explanation, writes the result back to the core, and returns the audit trail. The STP ratio is simply the share of submissions that reach a decision across these six stages without a human touch. Well-calibrated routing is what separates the risk that flows straight through from the risk that deserves a human pair of eyes, a topic covered in the guide to automatic underwriting routing and in the reference on the touchless underwriting pipeline.
What a good STP ratio looks like and how to raise it with an AI layer
A good STP ratio in insurance is not a single number. The STP ratio is the proportion of submissions that complete the underwriting journey with no manual underwriter intervention, over total submissions in the period. The calculation is direct: divide the number of submissions decided with no manual touch by the total submissions in the period. Two scope choices change the number materially and should therefore be stated. The first is what enters the numerator: only the auto-quoted business, or also the automatic declines and issuances, since a clean auto-decline still saved underwriting time. The second is the scope of the denominator: all submissions, or only those the insurer intended to automate, such as standard risks below a sum-insured threshold. A ratio over eligible submissions will be far higher than a ratio over everything. For that reason, an STP ratio is only comparable once the scope is defined, and a headline of 70% STP says little without knowing the line of business and the denominator.
The right target depends on risk homogeneity and data availability in the line. In high-volume, homogeneous personal lines, such as residential, simple auto, travel, and basic personal accident, high ranges are typically cited, often in the order of 60% to 90% or more for mature digital programs, because the risks are standardized and the data is structured. In small and medium enterprises and small commercial risks, the range typically cited sits in a middle band, in the order of 30% to 60%, depending on data quality and appetite breadth. In mid-market and large commercial risk, specialty, transport and cargo, and complex liability, a low STP ratio, typically well under 20%, is expected and correct, because each risk is heterogeneous and requires survey, judgment, or facultative reinsurance placement. These ranges are general industry context, not a regulator's statistic.
An STP ratio close to 100% on a heterogeneous commercial book is a warning sign, not a trophy. It can signal that appetite rules are too loose, that risks deserving scrutiny were bound automatically, or that the eligibility gate is letting through cases that will surface later as adverse loss experience. The objective is the highest STP ratio that holds loss ratio and risk quality within the appetite, not STP for its own sake. Alongside the STP ratio, it is worth watching the loss ratio on automatically decided business, referral accuracy, and the leakage rate, meaning in-appetite cases referred by mistake and out-of-appetite cases bound by mistake.
This is where an external AI layer comes in. An external AI layer is an intelligence layer that operates on top of the core and policy systems the insurer already has and connects via API. It reads the submission, enriches and scores the risk, applies the appetite rules, prices, routes the decision, and writes the result back to the core. It raises the STP ratio through three mechanisms. First, it automates the manual intake and triage, the document reading and the data extraction that previously forced every submission through a human, converting many in-appetite cases into straight-through cases. Second, it applies the insurer's own underwriting manual and risk appetite as machine-readable rules plus calibrated Machine Learning scoring, so eligible risks flow straight through and the rest escalate cleanly. Third, it keeps the core as the system of record, so the insurer modernizes the underwriting front end with no core migration. WIR operates as exactly this external AI layer.
Governance, explainability, and LGPD in the STP flow
Every automated underwriting decision must be explainable and auditable. For STP to be defensible, the insurer needs to reconstruct, for any case quoted, declined, or escalated automatically, which data was used, which rules and model outputs drove the decision, and why the case did or did not proceed straight through. Explainability matters above all on declines, which require a clear rationale the insurer can show to a regulator, a broker, or the insured.
Auditability is mandatory. A full audit trail at each pipeline stage serves both internal governance and SUSEP supervision of the Seguros e Danos market. LGPD, Brazil's general data protection law (Lei Geral de Proteção de Dados), governs the personal data used in scoring and in automated decisions. The framework requires a lawful basis, data minimization, and the data subject's right to request review of decisions taken solely by automated means. Encryption of data in transit and at rest, at every step, is expected.
There is also a deeper point. The models must reflect the insurer's appetite and manual, not a generic external rule set, so the decisions remain the insurer's own decisions. An AI layer calibrated to the risk-acceptance policy preserves the authorship of the decision and supports regulatory defensibility, rather than outsourcing judgment to an off-the-shelf model.
How WIR enables straight-through processing
WIR is the AI layer for insurance. An external AI platform that operates on top of the systems the insurer already runs, never in their place, and automates the quotation and underwriting journey according to the insurer's own risk-acceptance policy. WIR is 100% external, with no load on the insurer's IT and no core migration. It is not an insurer, a broker, or an MGA, and it does not carry risk. The core remains the system of record, and WIR writes the decision back into it with a full audit trail.
In practice, two modules drive the STP ratio gain. Underwriter Intelligence automates the quotation journey according to the insurer's risk policy, with real-time Machine Learning scoring calibrated to appetite, automatic routing by appetite and exposure, and predictive conversion analysis by product, risk, and broker, so the underwriter concentrates on risk and business development. Smart Sales works on the distribution side: it maps the portfolio by client and product, scores the next best action, and runs multi-channel campaigns with an attribution trail. Real-time dashboards, analytics, and reports give a proactive view of in-flight deals and pipeline. WIR's ML is calibrated to each insurer's risk appetite and underwriting manual, and every decision is explainable, auditable, encrypted at every step, and LGPD compliant.
Deployment does not flip everything to automatic on day one. Setup runs 3 to 12 months, covering automations, integrations, tests, and go-live adjustments, with clear scope and KPIs agreed before the start, followed by continuous operation in production after go-live. WIR was built with Mahway, a Venture Builder in California, and Avante, a Venture Studio in Brazil. The current public traction is a POC in execution with a global insurer in the Transport line. To size the opportunity in a given portfolio, it is worth pairing this read with the underwriting intelligence page and mapping which risks could already flow through STP. The AI layer for insurance. On top of the systems the insurer already runs, never in their place.
Frequently asked questions
What is straight-through processing (STP) in insurance?
Straight-through processing (STP) in insurance is the handling of a submission, from intake to a bound decision, with no manual touch by a human underwriter. In Seguros e Danos, a quotation request enters the pipeline, is read, scored, and priced, and is then quoted, declined, or issued automatically, because the risk is low complexity and sits inside the insurer's risk appetite. The defining feature is not speed. It is the absence of human intervention on the cases the insurer has already authorized rules and models to decide.
What is a good STP ratio in insurance?
There is no single good STP ratio. The right target depends on risk homogeneity and data availability in the line. High-volume personal lines typically cite high ranges, small commercial risks sit in a middle band, and large and specialty risk correctly shows a low ratio. These ranges are general industry context, not a regulator's statistic. A ratio close to 100% on a heterogeneous commercial book is a warning sign, not a trophy, since it can signal appetite rules that are too loose.
What is the difference between STP and a touchless underwriting pipeline?
A touchless underwriting pipeline is the design, the end-to-end capability to handle a submission with no human touch. The STP ratio is the scoreboard, the share of real submissions that actually completed without a touch in production. A line of business can be built to run touchless and still show a modest STP ratio, because many real submissions are still referred to a human underwriter for judgment. The pipeline is the capability, the ratio is the measured result.
Which risks can go through STP and which escalate to an underwriter?
Low-complexity risks that sit inside the appetite and underwriting manual can go through STP, with automatic quotation, decline, or issuance. The pipeline escalates out-of-appetite or high-complexity risk to an underwriter, and sends fraud signals or data gaps to investigation or enrichment. In WIR's Underwriter Intelligence, automatic routing operates by appetite and exposure, always with an explanation, freeing the underwriter for the cases that genuinely need judgment.
Does STP replace the insurer's core?
No. STP does not replace the core. WIR is an external AI layer that operates on top of the systems the insurer already runs and connects via API, never in their place. The core remains the system of record, and WIR writes the decision back into it with a full audit trail. The insurer modernizes the underwriting front end with no core migration, no load on its IT, and with decisions that are explainable, auditable, and LGPD compliant.