Transforming Insurance Underwriting with Google Cloud AI — Wrestling the Data Octopus
We discuss how Google’s Smart Analytics and AI technologies can help insurers achieve underwriting excellence.
In the first post of this series, we discussed the key building blocks that form the foundation for striking an optimal balance between fully automated, low touch and full touch insurance underwriting.
In the second post of this series, we discussed key challenges across the building blocks of insurance underwriting and how Google’s AI and analytics technologies rise up to the challenge.
In this post, we discuss how Google Cloud’s unique industry-focused technologies help insurers wrestle the proverbial data octopus to achieve underwriting excellence.
Intake Producer Submissions at Scale
While vanilla ingestion of data at scale is a no-brainer and has already been solved and proven by many cloud products, the real problem lies in not only ingesting but to extract meaningful data from various submitted forms, PDFs, images and other such ingested files, transform and store them in ready to use data sources — and all of the above happening in a single unified pipeline.
Ingesting producer submissions at scale with Google Cloud Pub/Sub also ensures,
- Scalable, in-order message delivery with pull and push modes
- Auto-scaling and auto-provisioning with support from zero to hundreds of GB/second
- Global message routing to simplify multi-region systems
- Cross-zone message replication and per-message receipt tracking ensures reliable delivery at any scale
And follow that up with transforming the submitted data in real time with Cloud Dataflow that allows for,
- Automated provisioning and management of processing resources
- Horizontal autoscaling of worker resources to maximize resource utilization
- Reliable and consistent exactly-once processing
Beyond transforming, you can also extract meaningful data from the various ingested documents using Google Document AI. You can automate and validate all your documents to streamline compliance workflows, reduce guesswork, and keep data accurate and compliant.
You can actually try it out here.
Geolocation data to enhance exposure management
Combining Google Maps, imagery data, mobile Android data and machine learning, we provide property attributes (residential, SME and corporate) for property underwriting risk assessment and existing property portfolio management. These attributes include all details about the property from geocoding to roof type, material, size, number of trees, building polygon, distances to nearest police, hospital etc all of which are among key pieces of data that underwriters can use for mature exposure analysis.
AI-Powered Data Warehouse and Data Science for Underwriting
Having an AI-powered data warehouse coupled with an advanced platform for collaborative data science can go a long way in improving underwriting decisions. Some of the other benefits include,
- Developing intelligent loss predictions to optimize underwriting policy pricing
- Organizing enterprise and sourced, structured and unstructured data to perform highly granular analytics to underwrite critical risks
- Forward-looking exposure predictions on potential future changes of current risk exposures to accurately assess propensity and viability of risk capacity
- Convert structured data into linked knowledge to quantify risk exposure, connecting companies, people, events, claims history, etc to boost underwriting efficiency, pricing accuracy and capture fraud
It’s important that underwriters have one unified platform to accomplish the above and what comes next.
Granular Analytics for Effective Pricing and Portfolio Steering
One unified platform for all your AI needs is also imperative to drive your pricing and portfolio strategy.
- Natural Language AI for advanced Entity Recognition and Entity Resolution for mapping coverage language into quantitative metrics
- Time Series modeling and analytics for a highly granular understanding of historical pricing variations vis-à-vis historical changes to risk exposures to underwrite most optimal and profitable pricing
- Develop models for active portfolio steering based geographical and other attributes
A global Insurance company, used Google Machine Learning to develop models to optimize pricing by predicting “large-loss” traffic accidents with 78% accuracy, enabling creating new insurance services such as real-time pricing at point of sale.
Google’s advanced collaborative data science platform: Vertex AI
We announced in the recently concluded Google I/O, our new Machine Learning platform called Vertex AI with industry leading capabilities.
- Build with the groundbreaking ML tools that power Google, developed by Google Research
- Deploy more models, faster, with 80% fewer lines code required for custom modeling
- Use MLOps tools to easily manage your data and models with confidence and repeat at scale
Vertex AI also includes pre-trained APIs for vision, video, natural language, and more — most of which may be used by underwriters as-is, to accelerate their analytics and decision making processes.
In this post, we discussed how Google’s AI and analytics platforms can help underwriters address the key building blocks we covered in the previous posts.
Do leave a comment if you’d like to know more or if you’d like me to delve deeper in future posts into any of the topics discussed in this series.
