Real use of chatbots and AI in the financial services industry

An (ideal) view into practice

A real use case from the financial sector: an "online consumer loan". A standard process that can be found on all financial institution websites.

In a digital world, applying for a loan online results in the loan offer being sent by e-mail. An included link to a personalised and interactive landing page that summarises the loan offer and allows the borrower to carry out interactions such as changing the loan amount or other modalities.

A chatbot offers the borrower its help to quickly answer any questions that may arise about the loan offer. A personalised interactive video explains the loan offer to the customer. Any change to the data on the landing page, within the chatbot or the video automatically results in a new loan offer, which the borrower receives via their preferred communication channel, e.g. email.

The next step in authenticating the person is carried out using a modern video identification procedure from a certified provider. The jump to the video identification procedure is also made from the borrower's personal landing page and is therefore fully integrated. The documents required for approval of the loan application (e.g. copy of identity card, salary statements, etc.) are captured via the photo scanning app integrated into the landing page and transmitted to the financial institution. After successful legitimisation and submission of all data and documents, the financial institution carries out a credit check in the back office and, ideally, approves the application. The borrower then receives an email with a confirmation letter for payment. This completely digital process eliminates media disruptions and long transport, waiting and idle times.

Use of chatbots depending on the objective and desired level of maturity

The CENIT chatbot maturity model is used in our consulting projects: We use this model to evaluate the respective customer situation in order to enable an appropriate process model and correlating product and technology selection based on the customer's requirements. The most common application scenarios are currently found in sales, marketing and service.

In sales, chatbots support customers in product selection and incorporate the customer knowledge they have already gathered. They focus on individual preferences or habitualised, historical habits in the purchasing or selection process. Rule-based technology is used here and usually operates with avatars. In marketing, for example, chatbots provide lifestyle recommendations, e.g. for the specific use of a product in an appropriate context. Here, too, there is little action outside of rule-based technologies, but the data sources that are used are numerous. In customer service, the picture is broader: simple customer enquiries are answered using rules. More complex enquiries must first be trained using machine learning and AI. This means that the full potential of the technology is not immediately available. Over time, however, the percentage of "suitable" answers to the questions asked is gradually increased. Machine learning is a mandatory prerequisite for "replacing" a real consultation. Here, the knowledge base and skills of the chatbot are optimised with controlled offline training. Only at the highest level of maturity is the chatbot able to optimise itself through continuous machine learning. In view of such solutions, companies are faced with the challenge of optimising and efficiently designing their multi- or omni-channel architecture for customer communication. Whether videos or chatbots are used: The aim is to be able to create a situation of personal interactivity in order to better serve and reach the customer. Unfortunately, many projects forget the human factor; technology is clearly assessable and selectable, but only secondary to the success of such automation. It is not the selected chatbot or the supposedly right software platform that is key, but the ideal flow of dialogue. So how do you motivate your project team to support the vision with top performance when, at first glance, the rationalisation of people is a benefit?

And the people?

In times of digitalisation, this applies to all areas of the company involved in the project and their people - from the technologies used and internal processes to the entire working attitude. Flexibility and the willingness to engage in lifelong learning are the most important qualifications that employees need to have today. This applies in both directions, so management must also realise that progress through the introduction of AI requires guidance and support. In particular, the use of state-of-the-art technologies, which trigger a series of change processes, is not always immediately accepted by all employees and must be accompanied by appropriate change management measures. The challenges of digitalisation, innovation and market pressure as well as often only rudimentarily described requirements coupled with constant change force an iterative, agile approach. It is not a question of slavishly adhering to all the requirements of the agile manifesto, but of adapting the method to the aspects that are useful for the project environment. The prerequisite here is that those involved have internalised the agile principles and are willing and able to take responsibility and work in a self-organised manner. The actual process model and the benefits of agile methods in the introduction of AI-based innovations must therefore be explained - to all those involved.

In our experience, trust and the enthusiasm of people are always the essential building blocks for the successful realisation of extensive, perhaps even difficult changes in customer communication - agile or not. We build on this and supplement it with AI-based technology.