Automation - Fraud Process

How to set up the Mohawk Automation (API) for streamlining the fraud process.

Mohawk Automation (API)

The Legentic Mohawk Automation (API) is our direct connection to Legentic data. This document gives you a few best practices. Our API is perfectly suited for automating standard processes. This documentation is written for a standard insurance claims process. The documentation will be applicable in other areas by extending the logic presented here.

This part of the documentation is concerned with automating the insurers' internal processes.

Claims Automation

Situation: you are trying to integrate with the Mohawk Automation in your fraud detection process.

Best practices: start with a single category (cars, for example) and build your automated processes from that position. We propose to run all incoming car fraud cases through our API endpoint.

Benefits: integrating in the fraud-process allows full insight into the automated results pulled out of the Mohawk Automation.

Investigation

Similar to the examples given in Claims Automation, the insurer can setup a system around Fraud Automation. This is described in detail there, and we'll only discuss what is special for the case of Fraud Investigation, namely the "deep link."

We've exposed a "deep link" that links back to the UI. For example, let's run our example query for a specific licence plate:

Python:

import requests

url = 'https://api.mohawkanalytics.com/classifiedads/gb/v1/license_plate'

headers = {'<YOUR_API_KEY>'}
data = {'registry_plate':'EN18UWW', 'limit':'20', 'select': 'mohawk_search'}

resp = requests.post(url, data=data, headers=headers)

print(resp.json())

Note that we have populated the 'select'='mohawk_search' parameter. This gives us the following response:

{'select': ['mohawk_search'],
 'count': 8,
 'documents': [{'fields': {}},
 ...
 'mohawk_search': 'https://app.legentic.com/investigate/results?query=volvo&country_code=ch&start_published_date=&end_published_date=&price_min=&price_max='

Following the link returned will (depending on whether you are already logged in, or still need to log in) directly to the Legentic Platform:

Implementation (limitations)

The deep link utility is implemented for all "simple" paths. Unfortunately, at this time the complex endpoints (/query and /free_text are not supported).

/license_plate
/vin
/phone
/email
/nickname

This solution allows the fraud investigator or SIU department to directly start investigating the item further and drill down on various selections in our UI.