AgentHack submission type
Enterprise Agents
Name
Balaji
Team name
Agentstronaut
Team members
Balaji.murugan@wns.com;rajeshwaran.ramaswamy@wns.com;archit.agrawal@wns.com;Varshini.m@wns.com
How many agents do you use
Multiple agents
Industry category in which use case would best fit in (Select up to 2 industries)
Insurance
Complexity level
Intermediate
Summary (abstract)
Auto insurance claims process involves several stages starting from the Insured filing the claim with details of Incident and bill estimates, this goes through inspection by a Insurance adjuster of all the documents and evidences submitted. Post approval the Claim is processed and the Insurer pays the full or partial amount to the Insured.
Detailed problem statement
The Challenge | The process is fraught with several challenges:
Coordinating communication, information sharing, and negotiations among policyholders, insurers, third-party claimants, repair shops, medical providers, legal representatives, and others can be challenging, especially in complex or high-value claims
Longer processing times, which can be frustrating for policyholders who are awaiting reimbursement
Misinterpretation of policy terms, inaccuracies in documentation, or human errors can result in claim denials or incorrect settlements.
The heavy reliance on human judgment of Surveyor introduces the possibility of bias and varying levels of competency.
Varying regulatory requirements and processes in each state
Detailed solution
The Solution | We have established an AI enabled automated solution for streamlining the Auto insurance claims process:
Email processing AI Agent - Help with classifying the customer queries related to claims, including inquiries about - claim status, needed documents, process steps, Coverages, situations like Thefts or Accidents, potential delays, Escalations and initiating appeals.
Surveyor insights AI Agent - Generate key insights required for the Surveyor to validate an auto insurance claim like - Compare the Narrative in Traffic authority report with Driver past history & summarise Driver violations in past. Check the Bill estimates are within coverages limit
HIL - Human-in-the-loop mechanisms when the Insured has expressed dissatisfaction or wants to raise an Appeal
UiPath IDP – For classification and extraction of data from Accord proposal forms, Traffic reports, Vehicle damage bills, Hospital bills, Driver history reports
UiPath RPA and GenAI – For analysing and summarising the snaps from the Accident spot. For automating the manual steps involved in the process
Demo Video
Expected impact of this automation
Timesavings of upto 80% for new claims request and questions regarding existing claims or escalations.
Accuracy in automatic estimation based on the damages to the vehicle will ensure the surveyor or adjuster has a possible range of estimates for the damages which would give leverage in adjusting the claim amount from the insurer or service personnel
UiPath products used (select up to 4 items)
UiPath Action Center
UiPath Agent Builder
UiPath Data Service
UiPath Document Understanding™
UiPath Maestro
UiPath Studio
UiPath Studio Web
Integration with external technologies
AWS bedrock, S3 bucket
Agentic solution architecture (file size up to 4 MB)
Sample inputs and outputs for solution execution
sample input content for email can be found in the below link -
Sample images for obtaining the estimation output -
Output from the estimator agent generated based on the above input images -
Other resources
Presentation -