AI Assisted Endorsments

AI Assisted Endorsments

Endorsement invoices sent to Clearmatch can be processed through our AI driven document extraction process which will extract the relevant data from the invoices and place on a web page in order for the details to be reviewed and altered where incorrect.

The users will also be able to view the eligibility check which is an automated check of the operational processes required to create an endorsement. Any checks that fail will be highlighted in orange to be reviewed and all checks that pass will be highlighted in green. 

The extraction of the data from the invoices and the eligibility checks will decrease the amount of time taken to process endorsements and also ensure consistency across all requests that are processed.

1. Navigate to the Clearmatch Admin Portal and select the "Endorsements " menu item under the "Clearway" section.



2. You will be presented with a page that will automatically list the endorsement requests in the last 90 days:






3. Use any of the filter options to refine the list presented

4. The small icon on the right of the menu will slide the menu back and forth to give you more viewing space on the screen.  




6. Each request will either have a  button or a  button on the right. "Assign" buttons appear if the request has not been clicked on before and no user is dealing with the request. The "View" button will appear once a request has been clicked on and assigned to a user.
Both these buttons will navigate users into the maintenance page  where the user can view all the details and make any changes.  If a request is unassigned, the first user to click on it will be assigned to the request. Once a request is assigned to a user, only that user and the operations team manager can make changes to the request or reassign or escalate the request to another user. This prevents more than one person actioning the same endorsement request.


Maintenance Page

Overview Section



1. When a request is received and the details extracted an attempt will be made to match the details in the invoice and email as best as possible to the actual values in Odyssey. Example
  1. Loan Number - if the loan number is not supplied in the email, the platform will attempt to match the "insured name" if supplied, and if not supplied the "client invoice name" to the legal and trading names of clients in Odyssey. If a match is found then the most recent active loan number for that client will be assigned to the request.
  2. Policy Underwriter - the platform will attempt to match the "underwriter" supplied on the invoice to the underwriters in Odyssey. If a match is found it will be used. If a direct match if not found, a fuzzy logic with AI learning match will be attempted.
  3. Policy Class- the platform will attempt to match the "class" supplied on the invoice to the insurance classes in Odyssey. If a match is found it will be used. If a direct match if not found, a fuzzy logic with AI learning match will be attempted.
Unfortunately the quality of the invoices supplied by brokers is very poor at times. This makes matching difficult.  To assist the user with reviewing the details extracted from the requests the following is done on each of the fields that need to be matched

  1. The matched values are presented on the page
  2. Underneath the matched value, the values that appeared on the invoice/email are presented.
    1. If presented in green it means a perfect match was found.
    2. If presented in orange it means a match has been made but the user needs to double check as it used fuzzy logic and AI to match.
This makes it really easy for users when reviewing as if green you do not have to look ay further.

Once the loan number has been matched, the details for the loan will be retrieved for Odyssey and will appear on the page. Details such as the interest rate, commission rate of the original loan will be loaded against the endorsement. The users will be able to modify these if they wish.

In addition and eligibility check will automatically be done. In this check the rules defined by the operations team will automatically be checked and as stated above, any rules that pass will be highlighted in green and any rules that fail and need review will be highlighted in orange. This will significantly speed up the time to process endorsements as only the checks in orange will need to be reviewed.

If the loan matched is incorrect, the user can enter the correct loan number and click "Reload" this will reload the page with the correct loan details and redo the eligibility check.

2. Any eligibility checks that fail and need review will need to be confirmed. Clicking the  confirm button  next to the rule will allows the user to enter a note detailing what they have checked and the reason for continuing. 

      

3. A request can be reassigned to another user in the operations team by clicking the  button on the "Overview" page".

4. A request by be escalated to another user in the operations team or the credit team by clicking the   button on the "Overview" page".


5. Notes can be added to the request at any time by clicking on . When adding a note documents can also be attached and uploaded.



6. Clicking   will save all the request details so they can be edited or processed later.

7. A request can be cancelled which will prevent it from being processed and uploaded into Odyssey, by clicking . Be careful as once cancelled, the request will no longer be able to be processed into Odyssey.

8. If the user changes certain details on the page which will affect any of the eligibility checks already performed, they will be informed that the change will result in the eligibility checks needing to be redone. In order to rerun the eligibility checks once all changes have been made , click the . This is only necessary if the changes made affect any of the checks, example the interest rate changes or policy amount changes, both of which affect the total endorsement value.

9. Once all the eligibility checks pass Or have been confirmed, the status of the request will be changed to "Ready To Process". At this point you will be able to process the request and import the endorsement into Odyssey. however if you realise you need to make a changes you can click the  button. 

9. Once either all the eligibility checks pass OR all the eligibility checks requiring review have been confirmed, the user will be able to  which will import all the details of the request into Odyssey. The following will be imported into an Odyssey endorsement quote

  1. All endorsement details
  2. All notes captured including notes against eligibility checks reviewed
  3. All invoice documents
  4. The original email will also be loaded as a document against the quote


Policies Section



10. Policy details can be reviewed and updated in the policy section. New policies can also be added or policies removed. 


Request Section


11. The request section will display the details of the original email sent , as well as contain the attached invoices.

12. The "Loan Settlements" section will display all the existing settlements on the loan, assisting users with their review.

13. The "All Client Loans " section will display all the loans the client ha ever had,  assisting users with their review.