Digital transformation requires changes in most activity areas of the company. It is not only about changes in IT systems, BUT…
Starting with a small project and delivering measurable results, is always a good way to efficiently prove to your organization the business case. I’d like to show you a case inspired by a real example, where a small IT project, lead to the digital transformation of the whole company.
Reading my case study will help you find the arguments to answer questions like:
- Can you ‘go digital’ without substantial investments in legacy systems?
- How to create a competitive advantage with workflow automation?
- How long do you need to wait to see the result of workflow implementation?
As is situation
An insurance company decided to improve the first notice of loss process (FNOL). They have already enabled its customers to notify about the claim via WWW page, but that’s it. The online form is used to capture the necessary information and as a result, an entry in JIRA like system is created. Next, all actions are executed manually by employees. There is some basic logic incorporated, that reminds the user about task due dates.
As the first step employee has to verify if the provided policy number exists in the policy system. Once it is confirmed, a claim is registered manually in the claim system and a confirmation letter is sent to the claimant. In case the claimant didn’t provide all the necessary documents at the notification stage, he is requested to complete them. The insurance company is waiting up to 90 days to get all the documents and either continue the case or close it. Once this step is completed, a decision is taken whether the claim should be evaluated by the technicians or a “fast track” claim handling process is applied.
Let’s digitalize and automate this process
Multiple systems are involved and the process itself may take some time to complete (due to waiting periods). This is a clear case of a structured workflow. It seemed obvious, that new solutions should include workflow functionalities. To take the final decision, that Camunda workflow and decision engine is the best option, the following requirements have been taken into account:
- Data entered once, should be used across all other systems
- Business should be able to adapt on the fly, the business logic that steers the process
- The solution should be highly scalable
- Core systems remain as they are
- Online monitoring should be available, and the data and parameters of all past processes’ instances should be available for reporting in external tools
Below I present a simplified FNOL process as a target model, designed in Camunda modeler. There are only 2 tasks (marked in orange), that require the manual intervention of the user. The first is to assure, provided documents are really what they are – in the future it is planned to use an OCR and AI algorithm for this purpose. The second is the waiting period for the remaining documents to be delivered.
The ultimate goal is to completely automate the FNOL process. The route to achieve this target state depends on several parameters. Having such a model in place you should judge:
- which steps generate currently the heaviest workload?
- which core systems provide interfaces?
- which route is the most frequent one?
- are there any legal or financial repercussions, if you fail to complete a step on time?
Having this in mind you can start with only one or two steps automated. Every next sprint you can automate another step. You can think about variants, where only selected cases will take the automated path and the rest will still go manual. A real example is the situation, where policies are managed in more than one system and one of them does not provide any interface.
This and other business logic can be easily steered by decision tables, that business can maintain themselves as shown below.
The FNOL process has been automated within 3 months, with 4 intermediary production releases. Once it has been done, the company continued to benefit from the workflow platform, within the claims area and further. The savings in the human work, the flexibility of business logic, the transparency of the processes, their KPI’s and the ability to monitor them on the fly, convinced other stakeholders to use a similar approach. Camunda has become the core for processes handling. New functionalities have been built with microservices approach and the existing ones are used where possible.
This shows that when considering the automation and digitalization in financial services, you should not limit yourself to standard options like substantial investments in your core systems or Robotic Process Automation with a focus on single step and not the whole process. There are other options, more suitable in given circumstances.
Generally, there are two messages:
- Introducing a workflow engine doesn’t mean you need to decommission legacy systems.
- If you are already busy with RPA, maybe it’s time to move on to the next level.