Role
Underwriter
What do you not know about a risk until the first claim arrives? Usually that the data in the application described the state before a refit, a move or a change in the line of business.
Intro call
Four questions we would start with
What do you not know about a risk until a loss happens?
If it is a lot, the question is not about the assessment but about what data reaches it at all.
How do you know the data in an application is current?
If the only source is a client's declaration from a year ago, the portfolio describes the past.
When is the decision about special terms made?
If it is in a negotiation, with no recorded reasoning, then a year later nobody will reconstruct why that policy looks different.
Whose desk does an unusual application end up on?
And how long before it comes back - because in quoting that usually decides whether the client stays.
We take over assembling the data, you keep the risk assessment
Taking in applications, assembling the data about a risk, an initial comparison against the criteria and portfolio review can be taken off the desk. Assessing the risk and deciding the terms stay with the underwriter.
- The application arrives with the data filled in. - What can be established from registers and from your own data is already with the application before the underwriter starts reading.
- A departure from the typical profile is pointed out. - Not as an assessment, only as a place to look at.
- The decision about terms is recorded with its reason. - At renewal and at a claim that is the information most sorely missed.
Scope
What can be taken off risk assessment
Eighteen example tasks. Each one has a human approval point. Below is a selection, and the scope we fit to the lines of business you write.
Taking in an application
- One route in for applications from agents and brokers
- Reading the data from the application and the attachments (verification: the underwriter)
- Pointing out gaps that make assessment impossible
- An initial classification of the application by product
Data about the risk
- Filling in data from public registers
- The client's claims history in your own portfolio
- Pointing out discrepancies between the application and your own data
- A signal when the client's line of business has changed
Assessment and premium calculation
- Comparing the application with the acceptance criteria
- Pointing out departures from the typical profile (assessment: the underwriter)
- Comparison material from similar risks in the portfolio
- A draft calculation under the tariff in force (decision: the underwriter)
Special terms
- A library of clauses and special terms
- Recording the reasoning for a departure from the standard (content: the underwriter)
- A register of the terms applied in similar cases
Portfolio review
- The structure of the portfolio by type of risk
- Claims experience by segment
- Pointing out risks departing from the assumptions (assessment: the underwriter)
A day's work
Today
- You retype and fill in the application data by hand
- You check the client's history in a separate system, if there is time
- You do not know whether a similar risk was already in the portfolio and how it behaved
- You set special terms without recording the reasoning
- An unusual application sits because it needs additional data gathered
- You review the portfolio once a quarter, in a spreadsheet
With AI assistance
- The data from the application and the registers is filled in before your assessment
- The client's claims history sits with the application
- Similar risks from the portfolio are pointed out as comparison material
- The reasoning for special terms is recorded with the decision
- An unusual application has its data gathered, so it waits less
- The structure and claims experience of the portfolio are available as you go
- Time to issue a quote
the quote goes back to the client sooner
the data is assembled before the assessment
- The quality of the portfolio
fewer risks accepted without the full picture
discrepancies and changes of profile are flagged
- Consistent assessment
similar risks are assessed similarly
comparison material is available
- Complete risk data
fewer decisions based on a declaration alone
your own and public data supplement the application
We start with the data at the application, because it decides both the quality of the assessment and its speed.
- Step 1
Reading the application and filling in the data
Data from the application, from registers and from your own portfolio in one place.
- Step 2
Discrepancies and changes of profile
Pointing out differences between the declaration and your own data.
- Step 3
Comparison material and calculation
Similar risks and a draft calculation under the tariff.
- Step 4
Portfolio review
Structure and claims experience by segment. Last, because it needs the risk data in order.
Common questions
Questions that come up at the first meeting
How is this different from a credit analyst?
Both roles assess future risk on past data. What differs is the subject and the regulation: there the risk of a liability not being repaid, here the risk of an insured event occurring. Different data, different periods and different consequences of error.
Will an underwriter be replaced by a rating engine?
A rating engine calculates, an underwriter assesses - and that distinction is the whole point here. Unusual risks, special terms and the decision to accept or decline need a judgement somebody answers for before the portfolio. We take over assembling the data about a risk, the work that leaves unusual applications sitting longest.
Does this fall under the AI Act?
That depends on the product, and it is the distinction we start every project from. Risk assessment and pricing in life and health insurance for individuals are treated in the regulations as a high-risk use, with obligations on human oversight, documentation and data quality. Property insurance for companies - what this entry describes - is not classified that way. The same item from this catalogue, moved to a life or health product, changes the classification of the whole - and that is why we set the scope against specific lines of business rather than against the role.
Still have a question that is not here? Book a call
Next step
How long does assembling the data to assess one application take?
We will show you what risk assessment looks like when an application arrives with the data filled in and the reasoning for the terms is recorded.