Telematics in auto insurance in Brazil turns driving-behavior data into an underwriting and pricing decision, replacing static proxies like vehicle model, postal code, and declared use with observed risk. The pull is structural. Auto (automóvel) is the single largest line of the Seguros e Danos (P&C) market, theft and repair-cost pressure keep loss ratios volatile, and the broad market grows at double digits per year while underwriting capacity does not keep pace. The competitive question for insurers is no longer whether to read the telematics signal, but how to score it fast and consistently on the systems they already run, with decisions that stay explainable and auditable.
State of the P&C insurance market
Brazil runs one of the largest insurance markets in Latin America, supervised by SUSEP, the federal authority for private insurance, and represented by CNseg, with FenSeg covering the property and casualty federation. The Seguros e Danos (P&C) perimeter covers auto (automóvel), property (patrimonial), transport (transportes), rural, engineering, financial lines, and liability, and the segment grows at double digits per year, according to CNseg reporting for recent years. Auto is the single largest line by premium (prêmio), ahead of property, with line-level premium and claims (sinistros) data published through SUSEP statistical dashboards.
Auto pricing in Brazil has historically leaned on static proxies. Vehicle model, postal code, declared use, driver age, and claims history correlate with risk only loosely. That matters for the loss ratio (sinistralidade), which in auto is highly sensitive to theft frequency and repair-cost inflation, both of which static rating tables react to slowly. Insurers should treat auto premium and loss-ratio magnitudes as figures to confirm against current SUSEP and CNseg statistics rather than fixed numbers. The structural relevance for telematics is direct. The same behavioral data that prices risk more accurately is also the data fragmentation that slows underwriting today, and that fragmentation is exactly what an external intelligence layer is built to consume.
What is pressuring underwriting
Five forces push telematics up the agenda while straining the underwriting (subscrição) function that has to absorb the growth. Theft and robbery (roubo e furto) are a top-tier loss driver in Brazilian auto, with severity concentrated by region and vehicle profile, which keeps premiums high in exposed segments and gives both drivers and insurers a direct incentive to price on observed risk rather than static proxies. Loss-ratio and pricing-lag pressure compound it, because static tables react slowly while behavior scoring lets pricing track risk faster and more granularly, within the actuarial and regulatory frame.
The operational strain has concrete numbers behind it. Underwriters spend around 40% of their time on administrative tasks rather than risk judgment, according to Deloitte. About 70% of insurers do not execute on innovation because of IT limitations, according to BCG, which is why a core rebuild is rarely the chosen path for a telematics program. Organizations also lose 20-30% of working time organizing unstructured data, according to Gartner, and high-frequency telematics streams add yet another data source on top of the heterogeneous PDFs, spreadsheets, and emails that brokers already send.
Distribution and data governance close the list. Insurance in Brazil is heavily intermediated through brokers (corretores), and 60%+ of brokers choose an insurer by response speed, according to Capgemini, so a behavior-based product only pays off if the insurer can score the signal and quote fast. Telematics data is also high-volume behavioral personal data governed by the LGPD (Lei 13.709/2018), which makes lawful basis, transparency, and consent design constraints from the start, not afterthoughts.
Risk, fraud, and the AI shift
The case for telematics-fed intelligence is a quality and risk-control case, not only a personalization case, and it carries its own governance burden. Where AI and Machine Learning enter the motor journey is specific. Ingestion and normalization turn high-frequency device and app data into usable inputs. Feature engineering converts raw traces into behavioral risk factors such as mileage, harsh braking, night driving, and road-type exposure. ML then scores each risk against the insurer's defined appetite and underwriting manual, rather than against a generic external score.
Fraud and anomaly detection move the control point onto the telematics stream itself. Behavior-based scoring narrows the gap between premium and risk that static proxies leave open, but it concentrates exposure to data quality, because a model fed by tampered devices or sparse traces misprices as badly as a stale rating table. The control is multi-factor anomaly scoring on both the submission and the stream, catching device tampering, trace spoofing, and mileage manipulation that no single rule would catch. ML also supports more granular pricing within the actuarial and regulatory frame, and it routes low-complexity in-appetite risks to straight-through quoting while sending complex cases to the right underwriter with the data pre-assembled.
The governance burden is real, and the responsible design answers it directly. Telematics is high-volume behavioral personal data, and decisions made on it must be explainable and auditable. Under the LGPD (Lei Geral de Proteção de Dados, Law 13.709/2018), data subjects can request review of decisions taken solely on automated processing, and processing must carry a lawful basis and transparency, as set out by the ANPD and the full text of the law. The credible model is an auditable one with a full decision trail, calibrated to a documented underwriting policy under SUSEP supervision, not a black box.
Where WIR fits
WIR is the AI layer for insurance. It is an external AI intelligence layer that sits on top of the insurer's existing core and policy-admin systems and automates the quotation and underwriting journey according to the insurer's own risk-acceptance policy. WIR never replaces the core, it runs no core migration, and it places no load on the insurer's IT team. The signature framing holds throughout. The AI layer for insurance. On top of the systems the insurer already runs, never in their place.
For a telematics program, that architecture is the practical path. The platform ingests the stream through multichannel intake, reads and enriches the context, and runs it through a risk and fraud engine that is a multi-factor ML model calibrated to appetite and the underwriting manual, then produces dynamic, risk-adjusted pricing and a decision that is a quote, an automatic decline, or an escalation to a human, always with an explanation written back to the policy core with a full audit trail. The products are concrete. Underwriter Intelligence automates the quotation journey with real-time ML scoring calibrated to appetite, automatic routing by appetite and exposure, and predictive conversion analysis by product, risk, and broker. Smart Sales maps the portfolio across client and product, scores upsell and next-best-action, and runs multi-channel campaigns with an attribution trail. Real-time dashboards and analytics give a proactive view of in-flight deals and pipeline. Every decision is explainable and returns a full audit trail, the data is LGPD compliant and encrypted at every step, and WIR is not an insurer, a broker, or an MGA, and carries no risk. Its only public traction is a POC in execution with a global insurer in the Transport line. Readers can compare this approach against a core rebuild in the WIR underwriting automation guides or start a conversation with the WIR team.
Outlook
Adoption is moving from pilots toward production in the lines where the data and the payoff are clearest, and auto is a natural starting point because the signal, driving behavior, maps directly to the loss, accident and theft, and the theft burden creates a built-in willingness to be tracked. Expect line-by-line rollout rather than a single large transformation. Given the cost and risk of core migrations, and the share of insurers blocked by legacy IT, the dominant architecture for incumbents will be external intelligence layers integrated with the core they already operate.
Two forces will shape the next few years. As SUSEP's Open Insurance framework matures, standardized and portable data should reduce the fragmentation that slows behavior-based scoring today and improve model inputs, as described in the SUSEP Open Insurance materials. In parallel, explainability, auditability, and LGPD-aligned consent and automated-decision controls are becoming table stakes rather than differentiators, and the winning models are the auditable ones calibrated to the insurer's appetite. Broker expectations on response speed will keep rising as digital and embedded channels expand, which makes faster, behavior-aware quoting a distribution-competitiveness investment, not only an efficiency one. None of this guarantees an outcome. It points to where the operational pressure and the regulatory frame are pushing the Brazilian auto market.
Frequently asked questions
What is telematics in auto insurance?
Telematics in auto insurance is the use of connected-vehicle and smartphone data, GPS traces, mileage, harsh braking and acceleration, and time of day, to assess motor risk on observed driving behavior instead of static proxies like vehicle model and postal code. It underpins usage-based and behavior-based products. The raw signal only creates value when AI and Machine Learning convert it into a risk score, a fraud check, and a price, calibrated to the insurer's own appetite and underwriting manual.
How does telematics data enter pricing and risk scoring?
Telematics data enters pricing and scoring through a pipeline. High-frequency device and app data is ingested and normalized, then feature engineering converts raw traces into behavioral factors such as mileage, harsh events, night driving, and road-type exposure. Machine Learning scores each risk against the insurer's defined appetite and underwriting manual, not a generic external score, and supports more granular premium calculation within the actuarial and regulatory frame. Behavior-based scoring narrows the gap between premium and risk that static rating tables leave open.
How does AI process telematics signals in real time?
AI processes telematics signals by ingesting and normalizing high-frequency device and app data, then scoring it against the insurer's appetite in real time. The same models run anomaly detection on the stream itself, surfacing device tampering, trace spoofing, and mileage manipulation that no single rule would catch. Low-complexity in-appetite risks route to straight-through quoting, while complex cases go to the right underwriter with the data pre-assembled. Every decision stays explainable and returns a full audit trail.
How do privacy and LGPD apply to driving-behavior data?
Telematics data is high-volume behavioral personal data, so it falls squarely under the LGPD (Lei 13.709/2018). Collection and use require a lawful basis, transparency, and purpose limitation, which makes consent a design constraint rather than an afterthought. Data subjects can request review of decisions taken solely on automated processing, as set out by the ANPD. The responsible design is therefore an auditable model with a full decision trail, calibrated to a documented underwriting policy, with data encrypted at every step.
Does WIR need to replace the core to use telematics?
No. WIR does not replace the core to use telematics. It is an external AI layer that sits on top of the insurer's existing core and policy-admin systems, ingests the telematics stream, scores it against the insurer's own risk policy, writes the decision back to the core, and returns an audit trail. There is no core migration and no load on the insurer's IT team. The Machine Learning is calibrated to each insurer's appetite, and every decision is explainable, auditable, and LGPD compliant.