Atidot

Atidot empowers the life insurance and annuity market to evaluate opportunities in an in-force book of business with big data and predictive technology. Atidot works with leading companies to discover under-insurance for up-sell and cross-sell opportunities as well as predict insurance behavior of customers such as lapse, surrender, option taking and premium persistency. Our technology is used by two core groups within the company: - Customer engagement opportunities are managed by either sales, marketing or distribution, as well as customer service. Atidot’s solution predicts customers’ probability to be under-insured, propensity to lapse or surrender and feeds the agent or the direct to consumer solution with the right insight regarding the customer. For example, providing the agent with leads from their book of business on specific customers with actionable insights for outreach. - Improvement of projection accuracy – Using Atidot, the company can generate much more accurate customer segmentation for financial projections resulting in reduction of volatility in projection, improvement of assumption, smarter segmentation for product development and pricing. Atidot’s solution sits on top of existing technology seamlessly. Our cloud based infrastructure uses API architecture that helps us integrate with core-systems, admin systems, CRMs and data warehouses and lakes. The solution is cloud based and is consumed as a SaaS solution, leveraging the scalability, redundancy, efficiency, security and privacy of the cloud environment. Our solution is based on a `Docker`/`Nix` container architecture in Azure, where pipelines of microservices form and run. Microservices include: ‘Atidot Data Tool’ for insurance-aware auto-tagging of semi-structured / raw data (using probabilistic algorithms and Deep-Neural-Networks), ‘Atidot Arena’ - running multiple machine-learning algorithms simultaneously in the search for the best predictor, ‘Atidot Storage’ - storage abstraction layer on top of CosmosDB, Blob storage, Data Lake. To orchestrate this, we have developed the `Atidot LIA` - a domain specific programming `Language for Insurance and Actuaries`. `LIA` (embedded in the `Haskell` programming language) lets actuaries, data-scientists and BI specialists program very concise actuarial scenarios. With `LIA`, teams focus on solving insurance problems, unconcerned by the numerous technicalities involved in data-science in practice (cloud, elasticity, devops, etc.). `LIA`'s underlying execution engines take care of that, firing containerized microservices in Azure. This unique technology is monetizing the vertical approach where insurance domain expertise is embedded to the model from the data discovery, based on proprietary insurance data layout, to the models that are built in an insurance specific way. Moreover, our LIA DSL provides industry level benchmarks for insurance behavior of policies based on the unique combination of internal and external data.