For companies in the finance and insurance industry, long-term data archiving is full of regulatory pitfalls. Artificial intelligence can support insurance companies in meeting the challenges of tedious data archiving and make them more efficient.
In connection with the growing flood of data in companies, keyword "Big Data," there is one aspect that is often rather hastily overlooked. And this is the direct and indirect costs of storing and archiving data. These costs can be quite considerable if legacy systems are kept running, especially for the purpose of archiving historical data.
The obligation to store data: from insurance agents to companies
There are many different legal obligations regarding the storage of data regarding digital processes. These obligations begin with the agent, who as a self-employed person, must comply with the GoBD rule ("Principles for properly maintaining and storing books, records and documents in electronic form and for data access"). This includes, for example, documents that were part of the initiation of a contract. These can also include emails which, according to the tax law, must be kept for a long period of time – and stored in a readable format(!) – for the tax authorities. Consultation minutes, policy amendments, change notifications, claims settlements – there is a long list of regulatory requirements that oblige companies to archive business transactions and documents on a long-term basis.
As is so often the case, the issue has both professional and technical aspects. The responsible department can take on the task of drawing up the rules for compliance. The German Insurance Association (GDV) provides support in this area, for example, with information and guidance.
When it is clear what data must be stored, then the technical details need to be worked out. There is also the question of format. PDF/A is often used as an ISO-standardized file format for many business documents. However, PDF/A should not be used without quality checks (for example, because of artifacts and unreadable document elements).
The GDPR, an additional pitfall
In addition to all the other legal regulations, archives must be GDPR-compliant. In the finance and insurance industry, companies frequently have to deal with documents that require special protection. For example, medical reports and health information. Protection against unauthorized access, the right of access to information, revision security and protection against tampering are just some of the parameters that must be taken into account when archiving.
Targeted archiving in legacy systems
One of the classic jokes told at the hotel bar in the evening is about the insurance consultants who discovered a server in the basement of their company, and that no employee knows what its purpose is anymore. The story is intended to illustrate how technologically behind the times insurance companies are.
Although the story always gets a laugh, it ignores the sheer necessity that leads to the continued operation of such legacy systems. The company's CIO is certainly aware that he/she is maintaining a data silo – with all the disadvantages and complications that this entails. But what is the smoothest possible way to transfer the old data to a new system, a DMS or an archiving solution? Many companies choose the supposedly most cost-effective option as their answer. They prefer to keep an outdated system running for the retention period.
However, this only appears to be cost-effective: the systems must then be supplied with updates, maintained and possibly equipped with new licenses for the entire period.
AI makes archiving more cost-effective
On closer inspection, it is often more economical to transfer only a part of the data stuck in a legacy system to a new system (unless the inventory is already completely outdated), for example, the data from the past three financial years. The other data can then be migrated to an archive where it can be evaluated. However, only the data and documents that actually need to be stored are actually stored.
Systems with artificial intelligence (machine learning) also demonstrate their strengths when it comes to data archiving. This applies both to initially setting up a data archive and to its subsequent operation. This simplifies the work of the system's administrators. Processed cases are automatically transferred to the archive. This saves administrative costs. The appropriate training of the machine components is necessary. More effective and flexible than rigid filter systems and rules, systems with artificial intelligence recognize and classify documents and send them to the archive with the correct status. And the outdated systems can finally be retired.
Our expert Jan Langkau will be happy to explain to you how AI can help you to manage documents. Make an appointment now.