Increase valid referrals of suspicious claims
Are you relying solely on adjusters to refer claims due to the high cost of an automated system that might not be as flexible as you like? Does your company use an antiquated, IT- controlled method for claims fraud referral? DataWalk can put you in control of how and why referrals are made, and enables you to extract pertinent information about ongoing trends.
Improve Effectiveness of Your Current Alerting System….
Quickly test hypotheses
Want to test complex fraud hypotheses that span many data silos? With DataWalk you can quickly blend “big” data from various silos – typically in days – and easily create and test sophisticated rules across the blended data set. Increase your understanding of your data – by seeing the connections - and start getting actionable results in just a week or two!
Link distant data to test for complex conditions
A foundational capability of DataWalk is to easily link data across distant sources, even if there is no common key. This can enable you to get new insights by finding relationships across data sets, with just a few clicks. For example, you could quickly associate insurance brokers with repair shops doing a high volume of expensive repairs, or connect weather conditions with accidents of expensive cars. DataWalk can generate connections not only for simple matches, but also for a variety of sophisticated conditions including geographic proximity, partially matching text strings, and many more.
Prototype your analysis and accelerate the analytics workflow
With DataWalk you can quickly prototype your analysis, to minimize risk and accelerate your analytics workflows. For example, DataWalk features a flexible logical data model (i.e., data structure) that is created using business constructs, so that a model can be incrementally, iteratively created and then easily modified without interrupting system operation. The model is created and validated in business terms, which eliminates the risk of “errors in translation” between business staff and technical teams. After prototyping the analysis, you can deploy the analysis using your production workflow, and deploy it much faster, while minimizing the risk of re-work in your process.
Easy for non-technical analysts
DataWalk is a highly visual system, which enables non-technical analysts to instantly ask and answer highly complex questions without relying on a specialist to generate a SQL query or write a program.
Dirty data? No problem!
Few companies have perfect “clean” data, but for DataWalk this is not necessarily a problem. As DataWalk is a connections-oriented system, multiple connections can often be identified between different records, such that the system often can still link data even if some values are missing or do not otherwise appear to match.
Combine alerts from multiple systems into an “alert hub”
If you have multiple alert systems in place, then DataWalk can enable you to easily create an “alert hub”, where all of the alerts can easily be brought together in one repository. With this you can easily join, split, and tune existing alerts to test outcomes on current and/or historical data, and you can identify new alerts to create based on such analyses.
Easily subset data for further analysis in R
DataWalk can also be used to easily subset data for further analysis, such as using R for predictive analytics, or Python for machine learning. Results from these tools can also be brought back into DataWalk to enrich objects or do further analysis.
…Or use DataWalk as an alerting system
Quickly deploy DataWalk as a powerful, flexible alerting system
For organizations that do not already have an alerting system, DataWalk provides a fast, powerful solution. The system can be deployed in weeks, and the highly flexible DataWalk system enables complex rules to easily be generated and tuned via a simple visual interface.
Powerful tools for exploring and investigating suspicious conditions
DataWalk includes various capabilities enabling you to quickly check alerts, and to do advanced investigations when needed. First, information about parties of interest is assembled on a dossier, which instantly presents relevant data from various internal and external systems, with the intent of enabling an alert to quickly be resolved. If needed, claims can be more extensively analyzed on network graphs, which visualize the connections between any desired people, places, and things. These objects can also be visualized on a map to instantly execute location-based analyses.