Root
risk pricing and segmentation from smartphone telematics
Root makes driving behavior measured by smartphone the primary variable in its auto pricing; in 2025 the app-first insurer reported 29% revenue growth and $40.3 million in net income.
Key points
- Auto pricing based on actual driving measured by smartphone sensors (2- to 4-week test drive).
- In-house ML risk-scoring models, retrained on claims history.
- Revenue growth +29% and net income $40.3M in 2025 (versus $30.9M in 2024).
- Evidence level A, confirmed status.
Objective
Set a fairer price than classic rating by measuring actual driving, screen out the riskiest segments, and improve technical profitability on an auto book.
The deployment
Root makes actual driving its main pricing variable. When a quote is opened, the app offers a two- to four-week test drive during which the smartphone's sensors measure braking, smoothness, cornering, time of day, and attention at the wheel. This data feeds a machine learning model that estimates individual risk and sets the price, where a classic insurer relies mostly on age, ZIP code, or credit score. Root describes this measurement as the strongest predictor of claims and the variable that weighs most in its underwriting model. Segmentation also serves to screen out a group of drivers judged too risky, up to twice as likely to be in an accident as the average target. In 2025, the company reported record annual results with 29% revenue growth and about 40 million dollars in net income (40.3 million versus 30.9 million in 2024), a sign that the data-driven model is delivering on its economic promise.
Results Proof A
2025 financial results published by Root, Inc. (earnings release and annual report Form 10-K filed with the SEC). The description of the telematics model and segmentation comes from the 10-K, an uninterpreted primary source.
How it works
Documented architectureThe stack in detail
- outil Application mobile Root (captation telematique) Two- to four-week test drive: the smartphone's sensors measure braking, smoothness, cornering, time of day, and attention at the wheel.
- outil Modeles ML de scoring et de tarification Root In-house machine learning that estimates individual risk from measured driving and sets the price; the heaviest variable in the underwriting model.
- infra Donnees de sinistres Root Claims history (outcome, cost) that feeds the continuous retraining of the risk models.
How it runs, concretely
For ops teams-
1Telematics test drive customer and AI
The prospect drives for two to four weeks with the app measuring their driving via the smartphone's sensors.
-
2Scoring and pricing AI
The machine learning model estimates individual risk and computes a personalized price.
-
3Segmentation and selection AI and actuarial
Segmentation screens out profiles judged too risky and targets careful drivers with a competitive rate.
-
4Monitoring and readjustment data team
Claims outcomes feed the retraining of the models and the adjustment of rates.
The measured driving behavior (braking, smoothness, attention, time of day). If sensor capture degrades or is not representative, the scoring and therefore the price lose reliability.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- Telematics driving data (mobile sensors or device)
- Claims history with outcome and cost
- Explicit consent to capture
Org prerequisites
- An insurer or MGA license
- An actuarial function to validate the models
- A compliance framework for automated decision-making and profiling (GDPR, AI Act)
Possible stack
- Custom telematics platform and SDK (the Root route)
- Third-party telematics SDK + in-house scoring model
- insurtech behavioral pricing solutions
The plan, step by step
- Step 1Frame with actuarial and compliance: permitted variables, consent to capture, the automated decision-making regime (GDPR art. 22 in the EU).Deliverable: Compliance file and validated variable scope.
- Step 2Set up telematics capture (third-party or in-house SDK) on a volunteer pay-how-you-drive program.Deliverable: App that collects reliable trips on a pilot segment.
- Step 3Train the scoring model on driving + claims history and backtest it with actuarial.Deliverable: Validated model with documented predictive lift over classic variables.
- Step 4Launch behavioral pricing as an option (good-driver discount) and clear product regulatory approval.Deliverable: Approved and marketed pilot product.
- Step 5Gradually raise the weight of the telematics variable in the main pricing, tracking the combined ratio.Deliverable: Book priced by driving with technical profitability monitoring.
First step: Launch an optional telematics program (pay-how-you-drive) on a volunteer segment before making it the primary pricing variable.
The same pattern, proven in other industries
Sources
- S1 Root, Inc. Announces 2025 Fourth Quarter and Full Year Results Primary archive pending
- S2 What is telematics technology - Root Insurance Primary archive pending
- S3 Root, Inc. Files Annual Report (Form 10-K) - telematics-driven auto insurance model Primary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.