|Predominately, there are two kinds of solutions in the market for fraud detection and prevention 1) Based on rule engine 2) based on AI. Its not possible to support the instant decision making by using the AI engine, since it can’t handle the real time data directly. For instant decision making, rule engines are used universally. For credit card or debit card transaction where response time is 8 to 10 second, such solutions are ideal. Though its possible to use combination of both to make more efficient and accurate prediction. We will discuss in this article different popular rule engine available in the market and combination of AI and rules solutions. There are five popular rule engines in the market:
FICO’s Blaze advisor
RedHat’s Decision manager
There is different way of comparing them in terms of price and ease of use.
|Rule engine||Price Range||Support cost||Ease of implementation|
|IBM’s ODM||100K GBP cloud version
400K GBP Enterprise version
|1000 GBP per day||Simple|
|200K GBP||1500 GBP per day||Easy|
|SAS||Same range as FICO||Same range as FICO||Easy|
|Redhat’s decision manager||24K for 16 core support
|Same as FICO||Simple|
|Drool||Free||Same as FICO
|Disclaimer on price: This price are approximate prices; you need to contact vendors for exact price.|
|These comparison gives an idea about the used solution in the market. Please check the link – (https://www.trustradius.com/business-rules-management-brms) for more information on rule engines. The decision to buy a rule engine is subjective depending upon relationship with different vendors.
Once we manage to decide that which engine, we can afford to buy, next task comes to implement them and integrate them with web application for application, transaction and claim processing.
Rule engine implements the rules generated by one of the sources:
Expert in Fraud: The current trends are that rules are generated by expert in fraud. There are different experts in different domain such as credit card fraud, vehicle finance fraud, auto insurance, health care and home insurance. Based on past investigation, they derive rules from different fraud cases and these rules are implemented in rule engines.
AI Engine: Company such FICO, SAS and Sparkling logic provides different applications for generating rules. Customer need to purchase these applications separately.
|AI Engine||Price Range||Support cost||Ease of rules Generation|
|FICO||Between 100K to 200K depending upon usage||800 GBP per day||Simple|
|Between 100K to 200K depending upon usage||800 GBP per day||Easy|
|Fraud Analyser**||Included in Fraud solution, free of cost||Included in Fraud solution, free of cost||Easy with Data Scientist support|
|Disclaimer on price: This price are approximate prices; you need to contact vendors for exact price.
**Fraud Analyser is our fraud solution.
|The following diagram shows the flow:|
AI Engine generate the rule
Expert generate the rule based on past frauds
Rules are implemented on Rule Engine
Rejected applications or claims are investigated, end user is notified of decision
If new rules are found, they are implemented in step 2
Rule engine is hosted on JBoss server
End user submit applications or claims, JBoss server replies with accept or reject decision
Web applications such as transaction, application and claims are integrated with Jboss Server
|Our team can support the rule engine integration and rule generation process at affordable prices. We have also support available for implementing the rule engine, integrating the rule engine with Jboss server and web application for transaction, application and claims.
In case, you chose our fraud solutions, your team will be working with dedicated data scientist to generate rules using our proprietary tools Fraud Analyser, key advantages are dynamic updates to the rule on more frequent basis at no extra cost, checking if current rules are sufficient to handle new way of doing frauds.