IFRS 17 is a beast with many heads. Starting from implementing the Contractual Service Margin (CSM) calculations to the impact on reporting: it affects the whole organization. The costs of the implementation are estimated 25 to 500 million euro per insurer. The CSM calculation requires every insurer to aggregate their contracts into groups, cohorts and portfolios. In order to do so all data of the contracts should be available in your system to make the correct judgement. The old dusty cabinet repeating history examples show there is still some way to go. This is no new information, however in my opinion insurers can make their own lives a lot easier using OCR or ICR while implementing IFRS 17.
An old dusty cabinet
Life insurers have (30 year) old contracts. Picture these contracts laying in a dusty old cabinet. Find yourself walking into that room with the cabinet. You open the drawer and after blowing the dust of the contract, you notice that these are all on paper. How are you going to aggregate these into cohorts, groups and portfolios? There is an option: make piles on the floor and start to puzzle!
Let us now imagine a situation without piles of hard copy contracts on the floor and go to a situation where insurers see the light and put data into systems already. As pointed out before data quality, and completeness are at the heart of a successful IFRS 17 implementation. IFRS 17 expects from insurers that they classify all contracts. One of the approaches is the full retrospective approach: envision going through your contracts; although having made the transition towards a more data minded approach, you probably still miss a lot of facts and figures in order to correctly value the contracts, because new requirements are rarely readily available in old contracts. So, you must head back to the old dusty cabinet and pull out the files to check for the missing data.
A brighter horizon
The analogy might come across as an exaggeration but gets the point across that data availability and quality are the preconditions for IFRS17 classification. So, in case of hardcopy files: how do you make an intelligent start at the inhuman task of putting all this data into the system?
Optical Character Recognition (OCR) or Intelligent Character Recognition (ICR) will help you out. Simply stated: OCR is a software program that scans through your document looking for the correct values and characters. The search configuration is an important step; tell the algorithm what to look for. It gives you all values you require in a standard format. However, you must keep in mind to configure which premiums you mean (premiums received upfront or paid afterwards and so on) If incorrect formulated, you end up with low data quality.
The techniques show promising results for insurance contracts. In short: you scan a document, save it, and let the software run the script. The results are the recognized values out the contract.
Is it that simple? In fact, yes.
There are three main advantages:
- Instead of feeding the data manually as a robot, let the robot do the work for you! Everybody hates these kinds of robotic iterations; contract after contract, resulting in high staff turnover and therefore mistakes are bound to be made.
- Reduce human error. Every human makes mistakes; a robot far less so. (unless you taught it wrong). So, you can have your people focus on higher scaled work like checking for anomalies and one-of-a-kind contracts. General out-of-the box OCR solution have typically a first-time-right percentage of 60%, dedicated training of algorithms can take this up to 85% within weeks.
- The much higher speed of these robots (depending on the contract up to 420x) reduces the run time of the whole process. Robot costs are on average 50% -90% lower than off-shore / on-shore employees
To sum it all up: with OCR or ICR the data is completer, less error and faster available. The time is now to replace that old dusty cabinet and start using OCR or ICR to kick start your IFRS 17 classification efforts.