How AI Redefining Mobile Banking: Best Customer Experience Practices

In September CactusSoft took part in 3rd Annual Conference Innovations on Banking in Prague, Czech Republic. A major focus at the Conference has been taken artificial intelligence followed by digital innovations, robotics in banking segment, customer authentication, chatbots and customer experience.

The COO at CactusSoft Mike Kartamyshev presented his speech on how AI redefining mobile banking and shared with the audience our best customer experience practices. In this article, you can read the most important points of his presentation.

Relatively recently companies in fintech and banking sector, in particular, received unique opportunity to build strong and positive customer experience through solutions and applications working on devices that became so popular and so addictive that we can easily experience panic attack if we forget them at home.

And if we’re talking about mobile banking software, customer experience, in this case, is built from the quality of the financial product offered and that’s something the application cannot influence. That’s where the client makes decisions, addicts to services he is provided with become an evangelist of the solution. And proper User Experience is vital here.

This is the last frontier where you can earn extra points in the eyes of your clients. Do not confuse it with UI, UX it’s a broader term that includes UI. So what do we do to provide this so-called proper UX?

Mike Kartamyshev
Mike Kartamyshev, COO at CactusSoft

To think in the right direction first we need to change our mindset. The Age of Advertising and Marketing is drifting away. People hate when they are pushed to buy. They do not trust promotions. We should not set a goal of selling or upselling as primary.

What people value and buy are attitude and care. Nobody wants to be offered a product or service at the moment they don’t need it. Instead, we should focus on providing a solution to clients’ problems and the satisfaction of their needs.

The ultimate goal with this mindset would then be to create an application that will make the client happier with doing just what he wants or needs at this particular moment. The perfect app would have one button that would know how to do the stuff I want right now.

Would anyone be able to complete this application? However, that is hardly possible to do at least at this time of technology development. Unfortunately or luckily we cannot read other people’s minds. We can already do a couple of things though.

mobile banking software

People are used to modern interfaces. They want to see features they use, to be not only functional but also beautiful and slick. It’s just a reality of modern world… unless you’re Craigslist then it’s your move.

Basically 80% of time clients perform following actions upon visit of mobile banking:

  1. Checking their Account balance
  2. Making Payments
  3. And Transfers
  4. Look at Transaction history
  5. And find Bank contacts to reach out to the support

What we can do then, is rethink the approach to the interface to make it convenient according to the client needs.

So-called Artificial Intelligence is used for problems that do not have an exact solution.

  • First, we collect data from users. Everything we can possibly get about them.
  • Then we clear, prepare and adjust that data. Math is needed here so you would definitely need some smart guys.
  • Then we choose and educate the model or algorithm which eventually, as a result, gives us a set of the probability of possible outcomes.

So for example, we want to implement a churn prediction model that will help us to reveal clients that will leave our bank within the following month. We collect client’s profile and behaviour parameters. We do know what clients left us in the previous month and that is smth that will help us train the algorithm. We prepare the data. Then using various tools and frameworks, we choose and educate appropriate model.

  • For that, we feed the AI with data prior to last month and let it adjust the models’ parameters to get a result that would be as close as possible to the actual list of clients that left the bank within that last month.
  • After the model is accurate enough we can use it to identify clients that have the highest probability of closing the bank contract within next month. Various types of algorithms and their combinations are used for prediction, segmentation, computer vision and natural language processing.
mobile banking artificial intelligence implementation

Just to remind, we’ve already improved the UI of the app which was the only the first step in keeping up with the modern world.

  • Let’s say we have a support team. In case the team has a well-defined script it can be applied in the chatbot which utilizes AI to transform the human question into a request machine would understand.

Not too much effort to have the first version. Yes, it’s very basic but it already helps in 80% of cases.

  • The next step would be to educate our AI to constantly monitor the client’s behaviour. And let’s say our AI understood the user recently spent some money on wedding stuff and now started spending on childcare items and that might be a good trigger to offer car loan for him to get an SUV or Estate car or maybe mortgage to get a bigger house.

Definitely that’s just a very simple example. AI is much smarter and will act in a much more sophisticated manner taking into account tens, and hundreds and thousands of parameters and building models that a human being would have never thought of.

The machine will figure out accurate patterns that trigger the suggestion of something the client with high probability might need at this time period. We transform ads into assistance in a particular situation that AI identifies.

mobile artificial intelligence implemented

With analyses of user’s behaviour and transactions, we can assist in reminding that a payment needs to be made.

Further development of bank AI ecosystem can allow us to implement suggestion for payments and transactions. For example, the system detects the client travelled a car and stopped in a parking lot in the city center.

If we enter the application we might want to make payment for the parking and the app will adapt its interface.  And even prefill all the required information so the client will only need to provide his fingerprint to confirm the transaction.

Eventually, banks can properly inform about any situation that requires user attention, help to improve the financial health and make financial forecasts to offer a range of required actions.

Do you remember the ultimate app with one button? We definitely are not at this stage yet but with proper approach and usage of these new beautiful technologies, we already can get closer to that goal.

To learn more about the topic or to see the full presentation, please contact us at

Share this page
If there is a project needing help or even a skill set you are missing, contact us.
Subscribe To Our Newsletter
To get new inspiring articles and news right in your inbox.