- Data is an essential tool to generate the high-value insights risk managers need to make confident and informed risk transfer decisions.
- Sophisticated tools that can facilitate automated data ingestion are needed to help produce these insights.
- Data ingestion is the process of bringing together large amounts of data into one place for convenient access and analysis.
- A smart process for data ingestion follows a few core steps: data capture, collation, review, presentation.
Data is an essential tool to generate the high-value insights risk managers need to make confident and informed risk transfer decisions. However, with such a vast amount of data to play with, sophisticated tools that can facilitate automated data ingestion are needed to help produce these insights.
But before we can understand what automated data ingestion is, we must first understand the principal concept – data ingestion. Data ingestion is the process of bringing together large amounts of data into one place for convenient access and analysis.
Typically, a smart process for data ingestion follows a few core steps:
- Data Capture
The first step of the data ingestion process involves capturing or extracting disorganised data – texts, images, documents – and converting it into readable and editable formats using optical character recognition (OCR) or similar tools. These tools then interpret the information and apply appropriate metadata for future analysis. They also standardise data formats, the first step in establishing and maintaining quality data.
Data capture in the Commercial and Speciality insurance market involves a large level of manual re-keying of data during the submission and quote process. This is often outsourced to third party providers, and is a slow process, with SLAs from 1 to 5 days for the return of a submission. With the inclusion of technology like automated data ingestion, complex multi-format submission data (Broker slips, Schedules of Values) can be extracted in minutes and effectively transform the underwriting journey for the better.
Collation follows, where the information is organised and linked to other relevant data in line with user or business needs. Text and data mining is part of this process. Similar to a search engine (like Google), data and text mining use different techniques to find certain predetermined pieces of information, whether based on rules or implied context.
This could include quote requests or asset schedule formats and files from within a business that can in turn help provide risk managers with better insights into areas such as asset acquisition and sale changes, geo-location data and current and past exposure data.
Once the data is collected and cleansed, it can be added to simple systems configured to support review, exception management and escalation, and requests for reviews and approvals within specific processes. Machine learning and advanced analytical techniques can enhance and improve decision-making.
For example, any quote requests that a non-approved broker submits would be flagged for further review or rejected outright. Similarly, coverage requests that involve high-risk regions, sanctioned markets or sensitive product categories would trigger alerts to the right users. Risk management teams or compliance teams would be alerted when risk appetite or contractual thresholds are breached or threatened.
With higher-quality data to analyse and more powerful technology for processing, bottlenecks are removed. Within risk assessment and pricing, using patterns found within historical data sets can support more informed decisions and provide customised recommendations for what to do next. Data that is presented clearly can help those users find the correct decision faster and more confidently, especially when presented directly within workflows or notifications.
Underwriting and beyond
Functions beyond underwriting can benefit from automated data ingestion processes. In sales, firms that digitise all quotes and proposals can learn from both bound and declined business, providing competitive market insights and enhancing business planning.
Exposure management, legal and compliance teams also stand to benefit. Compliance audits can be automated for increased accuracy and speed when all contracts are digitised. The same is true when legal liabilities need to be assessed, either broadly across lines of business or within narrowly defined product classes.
Ideally, an automated data ingestion platform will incorporate these capabilities within a single intuitive interface. Such a platform would harness the power of automation and AI to structure detailed information efficiently and integrate insights from structured and unstructured documents into standard data models. Using a flexible design, the platform will be able to meet the requirements of different formats and data structures while also complying with industry standards of clarity and transparency for risk definitions.