AI-Powered Insurance Risk Management Platform

AI done right.

Insurance
Assessment

The Insurance Assessment finds previously unidentified fraud, waste, and abuse, and shows the impact that the Daisy approach can have on your business.

Underwriting

The AI-Powered Underwriting solution evaluates 100% of claims data, aligning prices with risk, and driving higher profits while lowering operating costs.

Fraud
Detection

The Fraud Detection solution uses AI to augment data to discover collusion, non-obvious relationships, and outliers to dramatically lower false-positive rates and allow for more efficient investigation.

Claims
Automation

The Claims Automation solution identifies high-value fraud investigation and also segments claims for automation.

Loss Ratio
Optimization

The Loss Ratio Optimization solution finds the optimal combination of pricing, underwriting, fraud avoidance, and claim processing decisions to determine the desired loss ratio.

Daisy brings a unique engineering perspective to artificial intelligence and the insurance industry that dramatically improves business performance. Daisy’s Theory of Risk™ is a mathematical representation of the dynamics that generate profit in the insurance business.

We use optimal control as a reinforcement learning technique to identify anomalies in the system, such as risks and opportunities. Simulation is performed within the Theory of Risk™ to identify suspicious entities, including claims, individuals, physical and virtual places, and networks of individuals, as candidates to investigate, process and/or partially pay automatically.

Insurers can quickly detect and efficiently investigate fraudulent activity, saving them millions of dollars in claims payments.

Want to learn more?

Call 905-642-2629

Insurance Assessment

Daisy’s Insurance Assessment helps your organization find previously unidentified fraud, waste and abuse by identifying specific example cases. This assessment also delivers a business case that shows how AI could have impacted your business had it been in place in the past.

  • Daisy ingests your critical data including premiums, claims, policy, policyholder, claimant, payee, provider/supplier, vehicle/property and all other relevant data.
  • Daisy does not require ‘known bads’ to train predictive models for finding fraud.
  • Daisy shows you the range of fraud affecting your profitability, from low-volume, high-value complex scenarios to high-volume, low-value abusive behaviours, sometimes of equal or more value to the insurer.
  • Daisy deploys fuzzy rules, social network analysis, peer analysis, machine learning and previously existing analytical scoring, and aggregates them using fuzzy logic to create a Daisy Suspicion Index (DSI).
  • Daisy completes this assessment, delivers a business case, value reporting and our web-based investigations portal in less than two months from receipt of data.

Want to learn more about our Assessment solution?

Underwriting

Applying Daisy’s Fraud approach to underwriting data identifies anomalous applications. Combine this with the historical claims performance of peer or like applications using Daisy’s fuzzy logic, and the Theory of Risk™ creates a Daisy Risk index that can be used to create prices and approve applications. Daisy uses optimal control as a reinforcement learning approach to identify anomalies in the system (risks and opportunities), and simulation to assess risk and price more accurately. Daisy delivers a risk index and pricing recommendation with more accuracy using the following analytical techniques.

Daisy deploys fuzzy aggregation to combine all analytical methods into a single risk index for all entities (applications, people, and networks):

  • Pre-existing pricing rules combined with association rule mining on underwriting to create hundreds of rules that are implemented in a probabilistic form to permit ranking and risk prioritization of rule output.
  • Fuzzy logic framework incorporates existing machine learning model.
  • The Daisy system delivers predictive model risk scores that are incorporated into the fuzzy logic framework using peer application claims history.
  • Peer analysis compares all similar applications, people, things and networks using fuzzy logic to find anomalous behaviour and create a risk index. Behaviour is defined through thousands of features derived from underwriting data, provider, claim combinations, claim dates, claim details, VINs, diagnostic details and other data.
  • Social Network Analytics (SNA) resolves the identity of individuals that probabilistically share common personally identifying information. SNA also identifies communities of resolved individuals who are not in the same family group but are probabilistically connected through names, addresses, phone numbers, bank accounts, credit cards, social security numbers, social insurance numbers or any other attributes. SNA will link new applicants to existing policy holders with suspicious behaviour or present a high risk.

Want to learn more about our Underwriting solution?

Fraud Detection

Daisy’s Theory of Risk™ creates a Daisy Suspicion Index (DSI) rating for every entity in your data. The DSI recommends which claims, people and social networks should be investigated and which claims should not be processed. This DSI not only drives identification of high value, high certainty fraud investigations, but also identifies significant volumes of low value, high volume waste and abuse entities for review. Daisy uses optimal control as a reinforcement learning approach to identify anomalies in the system (risks and opportunities). Simulation is performed to find the best fraud decisions by optimizing recovery and avoidance. The low false positive rates Daisy achieves are the result of our DSI incorporating the following analytical techniques.

Daisy deploys fuzzy aggregation to combine all analytical methods into a single suspicion index for all entities (claims, people, and networks):

  • Pre-existing fraud rules combined with association rule mining to create hundreds of rules which are implemented in a probabilistic form to permit ranking and prioritization of rule output.
  • Fuzzy logic framework incorporates existing machine learning mode scores.
  • If ‘known bads’ exist, the Daisy system delivers predictive models incorporated into the fuzzy logic framework.
  • Peer analysis compares all similar people, things and networks using fuzzy logic to find anomalous behaviour and create a suspicion index. Behaviour is defined through thousands of features derived from provider, claim combinations, claim dates, claim details, VINs, diagnostic details and other data.
  • Social Network Analytics (SNA) resolves the identity of individuals that probabilistically share common personal identifying information. SNA also identifies communities of resolved individuals who are not in the same family group but are probabilistically connected through names, addresses, phone numbers, bank accounts, credit cards, social security numbers, social insurance numbers or any other attributes.

Want to learn more about our Fraud Detection solution?

Claims Automation

Daisy’s Theory of Risk™ creates a Daisy Suspicion Index (DSI) rating for every entity in your data. This DSI not only drives identification of high value, high certainty fraud investigations, but can also be used to segment claims automation.

  • Daisy can massively increase the volume of claims that can be straight-through processed, saving you time and money while speeding claims resolution and improving customer satisfaction.
  • Daisy identifies segments of claims that can be auto-adjudicated for partial payment.
  • Daisy also identifies segments of claims that require sampling review from less skilled investigators for payment denial.
  • Daisy will integrate these recommendations directly into the existing business process or Robotic Process Automation Systems.
  • All insights generated by Daisy are delivered from a single claim investigation and payments portal, and integrating all this data into one location saves time and money and streamlines the claims management process.

Want to learn more about our Claims Automation solution?

Loss Ratio Optimization

Daisy’s Theory of Risk™ is a set of mathematical equations representing the dynamics of insurance profitability. The year-over-year difference in company financial performance is related to the difference in quality of underwriting, difference in level of fraud, waste, and abuse avoidance, and difference in quality of adjudication decisions. Daisy’s control theory and reinforcement learning approach finds the optimal combination of pricing/underwriting, fraud avoidance and claim processing decisions to achieve any desired loss ratio. Daisy incorporates market competitive pricing benchmarks, investigator capacity constraints, auto-adjudication and straight-through processing thresholds as part of the Loss Ratio Optimization process to ensure the company achieves and exceeds it targets.

Want to learn more about our Loss Ratio Optimization solution?

Contact Us

Let us know how we can help.

We’d love to hear from you. Here’s how you can reach us.

TESTIMONIAL

One-third of the files opened in 2018 were as a result of information obtained from Daisy’s fraud detection system, accounting for 25% of Green Shield Canada’s total prevention and recoveries.

 

Tackling the Growing Problem of Insurance Fraud

Learn more how AI can help insurance companies save millions of dollars by detecting fraud faster and easier.