Order 4471: please confirm delivery date
Agents for the post room, where the work starts.
Monday morning in the post room. The scanner has been running since seven. On the third letter, the policy number is handwritten again. A claim with three phone photos sits next to a proposal that is missing the previous claims history, and both end up in the same queue. At an insurer, the work starts in the post room, long before a handler makes any decision. We believe an agent can take on exactly this part, so that your handlers finally get to spend their time on decisions.
Where your time goes today
Our post room matches hundreds of documents to policies every day, and often the policy number is only handwritten on the letter.
A minor claim for 180 euros sits in the same queue as a complex buildings claim, and in the end both take too long.
Proposals with missing information go back to the broker, but only days later, because nobody checked them for completeness when they came in.
Six processes an agent can take on here
Each one sets out what comes in, what the agent checks, what goes out and when a person has the final say. Your own process will have its quirks, and that is the one we want to hear about.
Register and allocate a claim
Check a commercial proposal for completeness and risk
Classify and route incoming post
Process a mid-term adjustment
Check an assessor's or repairer's invoice
Respond to a GDPR subject access request
Your process is not on the list?
Then it is probably the more interesting one. Use our Process Check to describe what piles up every week in three sentences, and we will tell you honestly whether an agent could take it on.
Classify and route incoming post
We start with incoming post, because that is where everything else begins. It is easy to check whether documents have been allocated correctly, and the agent gets to know your document types before it prepares any decisions.
What the agent needs
The agent needs policies and claims from the policy administration system, documents from the DMS and broker data from the CRM, plus access to every channel through which post reaches you. We bring this together in Microsoft Fabric and classify it with Purview, so health data and other personal data can be controlled separately and nobody sees more than their case requires.
Typical systems in your industry
What it looks like in practice
Questions insurers ask us
How do we make sure the agent never commits to paying a claim?
By not letting it commit to anything. The agent prepares and drafts. Settlement, policy issue and any statement to the customer with legal effect all go through human approval, without exception.
What about health data and the GDPR?
The data stays in your tenant in the EU. Purview labels the special categories of data, and the agent only gets the fields its case actually needs. Every access is logged, which also matters to your data protection officer.
How do we check the agent is making the right calls?
Every case has a log showing what came in, which checks were run and the reasoning. You can sample from it at any time. Many insurers start by letting the agent only make suggestions and measure how often it is right before giving it more to do. We think that is sensible.
Can this connect to our policy administration system from one of the big vendors?
Yes, through the interfaces that come with the system. The agent does not replace your policy administration system. It works in front of it and alongside it, and the policy system remains the source of truth.
What challenge is your business facing?
Your challenge is what interests us most. Send us a few lines about it. We will look at it free of charge, based on the systems you use today, and tell you honestly whether an agent is worth it.
