Behind the scenes: Developing the PremFina’s AI Chatbot.  - Questers

Behind the scenes: Developing the PremFina’s AI Chatbot.


Error message

Warning: Trying to access array offset on value of type null in _webform_render_mailchimp() (line 184 of /home/qstrscom/public_html/sites/all/modules/webform_mailchimp/

A talk with Kamen Kalchev, BackEnd Developer at team PremFina

Our partner PremFina has recently launched a brand-new AI chatbot on their website to offer even better client services and their team at Questers was heavily involved in the development process. We talked with our colleague Kamen Kalchev, a BackEnd Developer at PremFina's Sofia team, to share some details about the chatbot and his contribution to the project. Here’s what he shared:

1. You were actively involved in the development of PremFina’s new chatbot. Can you share some details about it? What is its main purpose?

I was very happy to work on the development of the AI chatbot alongside my colleague Hristomir Todorov, who is a back-end developer. The main purpose of the project was to lower the load our client services team is receiving from end customers each day. To achieve that, we wanted to build three main functionalities.

To start with, we wanted to build the functionality for the bot to answer frequently asked questions, which was the first and most simple but extremely useful one. The second functionality we wanted to introduce was related to the business side and was more complex to implement – the chatbot had to guide customers through a set of predefined questions and use those to amend customer details within our systems. Finally, we wanted to be sure that our customers are receiving the best service. Therefore, we had to ensure that the bot offers them the option to be redirected to our caring customer support advisors in case it fails to understand their messages, or they specifically ask to be redirected.

All that was a great challenge and we were very excited to work on such an interesting project.

2. You started developing the chatbot in April and two months later it was live on the official PremFina website. Tell us more about the development process and the tech stack you’ve used?

We kicked off the development of the chatbot on a three-day-long virtual hackathon with the main purpose to create a proof of concept of the chatbot idea.

We built Fina using the latest cutting-edge languages, modern databases and machine learning, natural language processing tools from Microsoft. Our bot is also fully integrated with PremFina’s systems and a third-party contact centre platform which our client services team uses for communication with customers.

The virtual hack finished with a presentation session where we had the chance to present our work to the other participating companies and the results were rather satisfying. This was followed by a few weeks of constant improvement of the chatbot capabilities after which we were finally ready to launch Fina on our website and customer portal.

3. Can you share more about your personal involvement in the project?

I was very happy to hear about the project at first since the AI field is something I am anyway interested in and did not hesitate to accept the invite to the virtual hackathon we started the development of the bot on.

We occasionally do internal hackathons where we have a wonderful time while learning a lot and getting a huge amount of work done. Due to that fact, we were all ready when it began, and we got obsessed with launching it very soon.

We like to keep the hack teams small so that we can be agile and fast. Therefore, sharing responsibilities to quickly resolve challenges becomes part and parcel of our hacks. I would qualify this as great teamwork. The project has surely taught me a lot and I hope that everyone else involved enjoyed it as well. I will be happy to have another hackathon with the team soon.

4. Did you have any bumps in the road? Can you share what was the most challenging part of this project?

There were some setbacks we faced in the process, of course, but we managed to achieve our goal with the cost of hard work and long hours, of which we are never afraid. It’s hard to identify the most challenging part. It was probably linked to the contact centre platform integration.

The platform our advisors were using at the time did not support third-party apps integration and we had to switch to another platform which supported that. This integration required both frontend and backend work, not to mention that customer support advisors had to undergo a mini training for the new platform too.

Having in mind that Fina would be working 24/7 and our advisors are only available in normal & extended business hours (depending on the requirements), we also had to make sure that customers could always reach us. Luckily, the solution to that was just a matter of integration with one of PremFina’s microservices and we were able to redirect customer requests to our advisors as an email, which are then processed as soon as they become available again.

5. Which part of your work are you mostly proud of?

Since we live in times of developing AI and automation, I am happy with the fact that we were able to handle some of the routine tasks within the chatbot so that our colleagues can focus on more interesting and creative ones.

We have an end-to-end reporting solution that displays real-time statistics on the usage of our AI chatbot. For the first two weeks after launching it, we were able to lower the load on our customer services team with 77.5% which enables us to scale even more without growing the team size that much.

6. Are you planning to add some new functionalities to the Fina chatbot?

We have actually never stopped developing new functionalities for the chatbot since it was launched. Our aim is to automate all the tasks that we can within a reasonable time.

Client services advisors have been very helpful too as they are constantly training Fina to learn more and respond/act appropriately. Naturally, the language recognition will get better and better over time as the training enhances.

7. Any other interesting projects you are currently working on?

Apart from the automation, we are doing for all departments in the company, there are a bunch of other interesting projects we are working on, including artificial intelligence, machine learning and blockchain projects.

We are building and maintaining microservices for the new infrastructure, which uses cutting-edge technologies. There are other ongoing projects in the AI field as well, and they are scheduled to go live in the near future.

For another view of the development process of the Fina chatbot, check out our interview with Hristomir Todorov, a BackEnd developer from the PremFina’s Sofia team who was also heavily involved in the project.

Further, PremFina’s team at Questers is growing and we are looking for more like-minded people to join their active and collaborative team. So, make sure you check out the hot vacancies.

Behind the scenes: Developing the PremFina’s AI Chatbot. | Questers


The website encountered an unexpected error. Please try again later.