Most transactions and processes that take place in companies involve some kind of document. However, 80% of these documents, due to their nature or format, cannot be processed as structured data. That is to say, although companies have them stored, they are not able to take advantage of the information they contain. And insurance companies are no exception. That is why today we want to tell you how Document AI can help the insurance sector. 

In the information age, businesses face the challenge of translating all the information that is created by - and to be consumed by - people into a language that can be processed by machines. What's more, they need these machines to be able to take advantage of that information. Document AI technology is addressing this obstacle.  

The digitalization of insurers

What is the digitized insurance consumer like? What are their interests and concerns? What are their needs? and how do they value the inclusion of new technological advances? An internal Google study provides answers to some of these questions.

Firstly, the study found that this type of consumer holds insurance companies' products and services in high esteem. In other words, they perceive insurance to be important and valuable, especially home and health insurance.

Another fact that stands out is that these profiles have incorporated digital transactions into their daily lives. Moreover, this is not a generational issue: 7 out of 10 consumers carry out this type of process. And this is true for all age groups. 

Finally, the study revealed that in the face of possible economic crisis scenarios, the digital experience is more valued by premium profiles. Therefore, those companies that improve their digital channels will start from an advantageous position when competing for this customer segment.

Benefits of Document AI for the insurance industry

The study conducted by Google not only focused on defining the profile of the digitized insurance consumer. It also investigated the characteristics of insurance companies and the sector in general. Two very relevant conclusions were drawn from this analysis:

• Insurance companies stand out for their ability to store documents and extract intelligence from structured data. 
• There are good tools on the market for analyzing structured data, but this is not the case for unstructured information.

In conclusion, insurance companies store valuable information and make the most of structured data. However, they lack tools for these other archived documents. This is where Document AI technology can strategically help insurance companies. Primarily, this improvement can occur in three areas:

User Experience

Implementing any of the solutions offered by Document AI allows you to improve the user experience in any of the processes that involve them. From onboarding to claims processing, from policy management to any other procedure. For example, instead of filling in multiple boxes on a form to provide your details to the company, this technology scans your ID card. 

Internal processes

Many of a company's internal processes involving documents can be automated with this technology. Document AI incorporates a double procedure:

• On one hand, the reading and extraction of information. 
• On the other hand, the processing to understand the meaning of the data, extracting specific fields and values and relating them, in order to take advantage of the automatic treatment of the data.

However, the true potential of Document AI becomes apparent when insurers establish their own use cases, delimiting specific problems and cases.

Data intelligence

Insurance companies have real "gold mines", powerful assets that they are not making the most of. Mails, policies, surveys, quotes... From all these documents valuable information can be extracted. 

So far, there was no technology available to do this. However, the main advantage of Document AI is that it can extract unstructured data that can then be processed by different management software.

Document AI for the insurance sector: some examples

Thus far we have learned about the ins and outs of the insurance industry and its customers. We have also explained the generic benefits of applying these AI tools in the insurance sector.

But to understand the benefits of Document AI in the insurance sector, we have selected three specific use cases. All of them share a double objective: from one point, to improve the user experience, facilitating the process. From another point, at an internal level, automating these processes will lead to operational cost savings.

The examples we are going to develop correspond to different moments in the journey that any customer may encounter in an insurance company: onboarding, a health questionnaire and a cancellation form. Let's look at them in more detail.  

Onboarding 

Imagine a scenario in which the user requests registration for a service through the online channel. By SMS or other instant messaging systems, a link could be sent to the user's mobile phone to complete the process. Through this link, the user will have to:

• Scan their ID card. Document AI will automatically extract every piece of information from that ID document: validity, nationality, place and date of birth, etc. This avoids the customer having to fill in a form with this information, making the procedure simpler and faster.
• Take a selfie with your phone camera. This photo will be used to carry out a biometric comparison of the user's identity. This step aims to avoid fraud and identity theft.

This example focused on onboarding can be extrapolated to any other process that requires the scanning of documents by the user: vehicle datasheet, driver's license, etc.

Health questionnaire

Another of the functionalities offered by Document AI is to train a Machine Learning model with its own forms. Let's give a concrete example: a person wants to take out a life or health insurance policy through a website or online channel. When requesting the assessment, the company sends them a form to fill in with their personal data and some questions related to previous episodes of illness, family history, operations...

When the potential customer fills in the data and the system receives it, thanks to having trained with AutoML its own model, it will be able to extract the data in two different blocks of information:

• On one side, it will locate the personal data provided by the user (name, ID number, date of birth, telephone number, address...).
• On the other side, from the questions and answers section, it will be able to separate the information required by the company to carry out the evaluation. 

The AutoML functionality, which can be combined with Document AI, is very useful for processing non-standardized documents. For example, it is recommended for forms or applications as these can change their format from one company to another. Thus, the company can train a specific Machine Learning model based on its forms.

Cancellation form

Although this is a scenario that companies do not want to see, another use case where Document AI can be very useful is in the processing of cancellations. 

Indeed, having this technology can be a competitive advantage. Its main virtue lies in the speed and efficiency with which it provides the data to carry out this procedure. As a result, the person in charge of this task has the necessary resources to counteract the customer's attempt to unsubscribe.

Here is a possible scenario for a cancellation request: the customer downloads a PDF form from the company's website that must be filled in and signed manually. The customer then scans it and sends it by e-mail to the department in charge of processing cancellations. In turn, the company agent who receives this request must either process the cancellation or prevent the customer from leaving.

All it takes is a simple query, a request via the app that processes the form in a matter of seconds and the agent will have all the data needed to handle the request in key/value format. This automation leads to cost savings by minimizing the time spent. Furthermore, the simplicity of the process and the quick response will be perceived positively by the user. Undoubtedly, an improvement in the customer experience can be another reason to stay with the company.

More Document AI Use Cases for the Insurance Industry

Artificial Intelligence applied to data extraction and processing opens up endless possibilities for companies. Beyond the use cases described above, there are a multitude of applications adapted to specific cases, forms and applications for each company.

Moreover, as we explained in a previous post, another advantage of Document AI is that it integrates with other Google Cloud Platform solutions, such as Dialogflow. This allows, for example, to design more functional virtual assistants for companies.