AI-Assisted Claims Handling, Starting with Cancellation Claims
Together with ERV, we are working on the claims handling of tomorrow and putting the vision of AI-assisted, intelligent claims handling into practice step by step. The first step is a document check for trip cancellation claims: the goal is for an AI to read the submitted documents, analyse the claim, check the facts and the costs against the insurance conditions and give the claims staff a well-founded recommendation for further handling. The final assessment of the claim remains with the staff, made much easier and focused on what matters.
- Client
- ERV
- Status
- In development
“What impressed us most was how quickly Voviva understood our requirements and delivered a first working version, and how well the team responds to our feedback.”
Davide SimoneHead of Claims, ERVAt a glance
Case prepared
Documents are sorted and read before anyone opens the case.
Every value traceable
Every extracted value is linked to the passage in the original document and highlighted there.
People decide
The system never decides whether a claim is covered; the staff do.
The challenge
In claims handling, a considerable share of processing time still goes into administrative work before the actual coverage decision: sorting documents, capturing the relevant information and requesting missing documents. Cancellation claims show this clearly. When a trip has to be cancelled, policyholders submit a stack of mixed documents, such as booking and cancellation confirmations, the policy and medical certificates. Staff sort these documents, filter and capture the relevant information, assess the facts and the costs against the insurance conditions and request missing documents where needed. Only then can the actual coverage decision be made.
The shared goal goes beyond a single type of claim: claims handling in which the AI takes over the time-consuming administrative groundwork before the case is opened, but leaves the decision on the claim to the staff. Cancellation claims are the first step towards that.
Our solution
The solution is in development and is currently being tested with ERV. The first step is a document check for cancellation claims. It takes in a case from the online form or by email, splits combined PDFs into individual documents and categorises them on its own. It then extracts the relevant information, with every item linked to the corresponding passage in the original document and shown there.
The AI checks the facts, summarises them in a traceable way and verifies the costs against the insurance conditions; for simple fee models, a fixed rule recalculates them. The policy is looked up automatically and the system proposes a recommendation. Missing documents are listed directly in a draft request.
The document check is one of several steps. Alongside it, we are developing the bigger vision of AI-assisted claims handling further with ERV and defining the next stages of expansion. The building blocks of the document check, from splitting and classifying the documents to traceable extraction and the check against insurance conditions, are built so that they carry over to other types of claim.