The creation of the chatbot everyone has been waiting for, for the insurance company Vitaminise

How did we achieve this result?

SendPulse is an omnichannel marketing platform specializing in creating chatbots, web push notifications, landing pages, email newsletters, and much more. They provided active technical support and consulting, as well as increased the number of variables for the correct operation of the bot. 

The Diia application, the databases of the Ministry of Internal Affairs and the Motor Insurance Bureau were a cooperation for quick access for users to the necessary data. 

All this was done to simplify the process of issuing and purchasing insurance policies, with maximum automation of the process, specifically to reduce human involvement in filling out forms. 

Problems

Users had to go through a difficult bureaucratic process: 

• A large number of documents need to be found
• Transfer data to the insurance company before seeing the price for insurance.

In the bot, he only needs to provide data about the car, and the price is calculated automatically, all data is at hand.

Our product was designed to eliminate such problems:

• make the process of purchasing a policy clear;
• make the product accessible through omnichannel. Omnichannel allows for convenient and efficient access to necessary information from the most user-friendly sources.
• order a policy at any time;
• reduce the time spent by the client on policy registration through automation;
• set up integration with Diia for quick data retrieval when filling in the relevant fields with information in the policy;
• improve lead generation and increase sales levels.

Features that arose during the construction of the policy acquisition logic:

• the policy start date cannot be earlier than tomorrow
• the policy can be issued by individuals aged 16 and older
• the insurance company validates the city through the COATUU code: the user could search for the city either by name or by index. Some data was outdated and had to be updated.

Solution

To achieve the goals set before us, it was decided to develop a chatbot. 

Chatbots are modern assistants in automating marketing processes. With clearly defined algorithms, most responsibilities of live personnel can be delegated.

The advantage of chatbots is:

• Availability 24/7;
• Instant response to basic questions;
• Improved service quality through data collection process automation;
• Personalization of communication;
• Reduction of customer service costs.

Functions

Functions that can already be tested:

• Convenient auxiliary options, such as reminders about policy expiration or the ability to renew a policy with one click;
• Accelerated insurance processing;
• Monitoring of existing insurance policies: verification, tracking, and providing up-to-date information.

Customer Journey

Work Process

The purchasing process begins with obtaining data from the user:

• we check for previous searches and already filled data
• in the action, you will need to authorize additionally, as the access term to the user’s account quickly expires

Next, the user is offered to choose a method for searching for a vehicle:

The vehicle is searched in Diia – then by phone number, and only if the number is also not in the database, it is entered manually.

In the Sendpulse constructor functionality, options such as API requests are available. Through this function, our team was able to enhance the chains, expanding their capabilities. This allowed us to set up integration with Diia and the ability to search for vehicles in the MTIBU database.

The vehicle search is carried out both through Diia thanks to the API request to your personal account.

All data will have to be entered manually.

This is the initial validation

Issuing information on all parameters of the vehicle:

Manually – if used 2 methods – entered manually

If there is an error, you can search for the vehicle again.

1.40 – if nothing was found

There is an option to search for a specific type of vehicle. When you fill in the data manually, you will need to remember the necessary vehicle parameters.

For each type of vehicle, we took into account all the nuances when filling in: for example, the difference in classifications of engine types.

When entering incorrect data, you can manually correct a specific parameter separately without having to re-enter all the data again.

Data validation is carried out based on a number of parameters

If there is a lack of data, the bot will request it

Additional vehicle validation stage

There were difficulties in finding the registration location, as the data could not be correctly retrieved 

We solved this problem by allowing the user to specify the postal code of the city and then providing the result.

When issuing a vehicle, the number of vehicles owned by the user is also taken into account. There is an option to choose which car they will insure.

The cost calculation is based on a complex economic-mathematical model that takes into account the specified parameters.

The data is sent to the company’s service

Initial and final calculations

If the user is satisfied with the price, the next step is to arrange the service.

For this, user data is required, which can be obtained either automatically or by entering manually

When issuing the policy, the user’s age and the possible presence of an old-format passport are taken into account.

At each stage, there is a verification node and the possibility to correct the data

Also, the start date of the policy cannot be earlier than the next day

The next step is the validation of the contract creation

Payment verification

In the end, if everything goes well, the policy is sent immediately to the email

SendPulse has certain limitations, so changing the data manually is not optional.

We have developed a logic that allows the user to change personal data without having to refill all fields. (see the screenshot below)

In the bot, you can view all data on existing policies (validity period, link to the document, technical specifications under which the policy is registered, policy number, and product type)

The accuracy of the data is confirmed through requests in Diia.

The developed algorithm for searching vehicles can be performed both manually and automatically through the Diia application, quickly verifying the accuracy of the information

Personalization of the bot

Even in the welcome series, we took into account the smallest details: if the user registered earlier, the bot addresses the user by the name previously provided.

In case of an error, information about its type is provided.

With each new cycle of purchasing a policy, we take into account the presence of previously provided data.

With each change in business logic, we can easily implement it into the bot’s logic, as we maximize the separation of all stages into sub-stages.

Result

During the project, an omnichannel chatbot system was developed for social networks such as Telegram, Viber, WhatsApp, and Instagram, with the main idea being the automation of all processes when purchasing an insurance policy. 

A unique logic for the operation of chatbots was developed, including data conversion, validation chains divided into stages, and data correctness checks. 

Additionally, the company manually expanded the database of city codes (postal codes) for accurate completion of the user’s registration address.

When using the chatbots, the user can request data from the Diia service for automatic completion of all fields. At the same time, the user has the option to manually edit the necessary fields. 

Each bot is interconnected, as when purchasing through one social network, the user can easily find their account on other relevant platforms and check existing policies / purchase a policy using the phone number that was previously entered into the database.