top of page
Frequently asked questions
We understand that when considering our solutions, you may have questions about how we work, what we offer, and the kind of support you can expect. To make things easier, we’ve compiled answers to the most frequently asked questions from our clients. This section is designed to provide you with clear, straightforward information so you can feel confident and well-informed as you explore our services. If you don’t find the answer you’re looking for, our team is always ready to assist you directly.
Company & PositioningSherpa Plus Platform OverviewSherpa Lens - AI OptimizationSherpa Engage - Retention & Loyalty LayerAPI Integration & Operating FlowsUnderwriting Risk Control & Claims PhilosophyPayments & Crypto RailsCompliance Security & ReliabilityEcosystem Orchestration - Beyond TechRoadmap & Agentic AIMarkets Regions & Use CasesPartnershipsBusiness Model & EconomicsCompany Team & CredibilityOnboarding & Partner JourneyComparison Q&AsAI & Optimization AdvantagesAPI & Architecture AdvantagesMGA Capacity Structuring AdvantagesSpeed-to-Launch AdvantagesProduct Co-Creation AdvantagesRegulatory & Multi-Market AdvantagesClaims & Full-Stack Operating AdvantagesDual-Side Enablement & Ecosystem Advantages
What is Sherpa+Lens?
Sherpa+Lens is Gangkhar’s real-time AI optimization engine that continuously improves pricing, messaging, and product versions based on data and behavior.
What outcomes does Sherpa+Lens optimize?
It targets higher conversion and attachment rates, higher ARPU, improved retention and loyalty, and better revenue per lead—while managing risk performance.
How does Sherpa+Lens work in real time?
It ingests live signals from policies, claims, and customer behavior, then dynamically adjusts offer structure, pricing, and messaging.
What kind of AI techniques are used?
Lens references multi-armed bandits and goal-based AI, with specific methods like Thompson Sampling and LinUCB for experimentation and optimization.
What machine learning models are referenced?
Sherpa+Lens references Gaussian Mixture Models for clustering/segmentation and classifiers such as LightGBM, CatBoost, and XGBoost.
Does Lens support automated experimentation?
Yes. It includes an “Experimenter” capability for automated tests, designed to reduce or avoid seasonal bias.
What is the “Recommender” in Lens?
It’s a capability that enables more personalized product recommendations based on segment and behavior patterns.
What is the “Price Optimizer” in Lens?
It’s a capability that identifies optimal prices for segments using large-scale simulations.
How does Lens avoid “one-size-fits-all” pricing?
Through segmentation and real-time learning, Lens can vary pricing and offers by segment and observed behaviors.
Is Lens designed for high-volume businesses?
Yes. It’s suited for businesses with high transaction volumes (e.g., per-ride, per-purchase, per-session) where continuous optimization matters.
How does Sherpa+Lens ensure safe deployment of AI models in a regulated insurance environment?
Sherpa+Lens follows a governed AI deployment path: models first run in shadow mode (observing without affecting decisions), then progress through canary stages (10%, 30%, 60% of traffic) before full rollout. Every stage includes human-in-the-loop validation, ensuring that no model goes live without oversight. This approach is critical for regulated insurance environments where explainability and control are non-negotiable.
How does Sherpa+Lens handle model versioning and performance monitoring?
Sherpa+Lens includes full model versioning and feature drift tracking. Every model version is recorded and traceable, and the system continuously monitors for feature drift—changes in input data patterns that could degrade model performance. When drift is detected, the system flags it for review, ensuring models remain accurate and reliable over time.
What future capabilities are on the Sherpa+Lens roadmap?
Sherpa+Lens has several capabilities prepared for future integration, including churn prediction (identifying customers likely to cancel), fraud detection, advanced product recommendations, and coverage personalization. These are designed to extend the optimization layer beyond pricing and messaging into deeper lifecycle management.
bottom of page