IBB AI ChatConsultant
The project was developed as part of the Lean Startup Sprint, a collaborative initiative by Berlin Startup Incubator and Investitionsbank Berlin, a public bank in Berlin dedicated to providing financial support to entrepreneurs at all stages of business development.
SITUATION
IBB Bank, like many German institutions, has faced many difficulties with digitization. Untrained staff, lack of proper software and procedures mean that day-to-day work takes consultants much longer than it should.
TASK
The task we faced was to use AI technology in the bank’s day-to-day operations to streamline the consultants’ work and improve their contact with customers.
ACTION
We started our work on the project with in-depth interviews with employees of the IBB bank and other similar institutions in the Berlin market for financial support programs for startups, in order to learn how they work and identify their difficulties.
During the research, we learned that while 60% of the customers of these institutions prefer online contact, this number among bank employees is only 28%, which is a big discrepancy.
The lack of willingness to use technology and the low degree of digitization of the bank are the reasons for the difficulty of having to sort emails manually, and therefore time-consuming, between departments.
The lack of information flow between departments in electronic form is also a problem — the only digital tool the bank uses is Outlook, and some consultants print out emails and continue their work in analog form anyway.
Consultants face a lack of trust from customers (mainly among imigrants) and customers face a language barrier.
Very often turns out that customers do not know what they need and are not familiar, even to a basic degree, with the bank’s offerings.
SOLUTION
The solution we propose will relieve consultants of the burden of their daily work and give customers the tools to better educate themselves about the bank’s offerings and their options.
The chatbot, based on LLM, will be the first contact customers have with the bank. With its help, you can check the subsidy program suitable for your situation, learn about its details, submit an application, send a message to a consultant or make an appointment to see him or ask any question.
The AI model trained with the bank’s offerings, departmental structure, special vocabulary and hundreds of times of former customers will be able to identify users’ problems well and offer them the right solutions.
As the chat conversation proceeds, the tool will collect all the necessary data about the user and organize it to fit the format of the contact form visible on the left side of the screen.
In case the answers provided by the customer are not very precise, which would not allow the consultant to understand the customer’s business situation, Chat will ask additional in-depth questions making sure it collects all the necessary information.
At the end of the chat, the customer will receive suggestions for programs that are suitable for him, can send a message to the consultant with a redirect to the appropriate department and book an appointment with the bank.
SUM UP
A chatbot is in no way intended to replace a live relationship with consultants. It is a tool to help as the first contact with the client that does not require any work on the part of the consultants, so its success doesn’t depend on the consultants’ ability to use digital tools.
With Chatbot, the customer is well informed, and the consultant is assured of getting all the necessary information to help the customer, plus no one has to manually sort emails between departments.
NEXT STEPS
Another important step in this project would be to connect the Chatbot to the bank’s CRM system. The LLM model having such a large amount of information from the bank’s scope of work would also be able to help in the process of verifying financing applications.
