Smart digital interfaces are the new horizons of UX

 

Online banking platforms are increasingly incorporating built-in accounting tools and other products tailored for entrepreneurs. Mobile banking applications, once perceived as supplementary, have evolved into full-fledged services, enabling business owners to manage key tasks like account operations directly from their devices.

The Next Frontier in User Experience

What’s the next step in enhancing the user experience? How can remote banking interactions become even more seamless? What breakthrough will redefine banking interfaces? My prediction: we are rapidly approaching the era of intelligent interfaces. In the near future, the work of analysts and designers will be complemented by artificial intelligence capable of analyzing and customizing banking interfaces for each individual among hundreds of thousands of clients.

Evolution of Banking Apps

By the end of 2017, the capabilities of mobile banking applications had grown significantly. Previously, they seemed like incomplete extensions of online banking. By then, users could execute payments, download statements, and manage cards. Today, leading mobile business banks allow payroll management, foreign currency control, and other complex tasks. This progress has led to a growing number of mobile-only users among business clients.

Similarly, online banking for businesses has evolved, introducing tools like document workflow management, built-in accounting for small businesses, and ecosystem services tailored to entrepreneurs.

Challenges in Expanding Functionality

While the functionality of banking apps continues to grow, integrating a comprehensive accounting system for general taxation or a robust inventory management system remains challenging. These require solutions tailored to various business needs and operations, which can complicate integration.

---

Personalized Interfaces: A Game Changer

Imagine a personal banking consultant who is fully competent in financial products, aware of your unique context, and adept at anticipating your needs. This is the promise of intelligent interfaces—a system that learns from the user’s behavior and adapts accordingly.

Key features of such interfaces:

1. Shortcut Creation: Routines like recurring payments can be completed in just a click. For instance, the system can automatically prepare a payment for June if the last one was for May—no templates or manual input required.
2. Dynamic Navigation: Frequent tasks like payroll management appear prominently, while rarely used features like currency control are tucked away but accessible.
3. Proactive Support: Interfaces will preemptively generate filtered reports or highlight upcoming deadlines based on user behavior.

---

Current Examples and Future Directions

The concept of intelligent interfaces isn't limited to banking. Early examples can be seen across industries:

• Text prediction in messaging apps suggests frequently used phrases.
• E-commerce platforms like RichRelevance personalize searches and recommendations using machine learning.
• Chatbots in banking offer guidance on financial products, with potential expansion into areas like credit advisory and currency control.

The next step involves combining text-based and voice interfaces to create assistants capable of tasks like opening accounts, issuing cards, or suggesting better plans—all tailored to the user’s context.

---

Overcoming Implementation Challenges

For intelligent interfaces to thrive, banks must unify all client interaction data—online sessions, branch visits, ATM transactions, call center records, and more. However, this poses significant challenges due to:

1. Data Fragmentation: Consolidating information from disparate storage systems.
2. Infrastructure Overhaul: Upgrading legacy systems to accommodate AI-driven solutions.
3. Legal and Security Compliance: Ensuring privacy and security across integrated platforms.

---

Technology Underpinning Intelligent Interfaces

Two key technologies driving intelligent banking interfaces are:

1. Recurrent Neural Networks (RNNs): Useful for analyzing sequential data like user behavior logs. Advanced architectures like LSTM (Long Short-Term Memory) can identify long-term dependencies and suggest shortcuts for recurring tasks.
2. Reinforcement Learning: This approach enables systems to learn through interaction and feedback. For example, interfaces can balance between offering frequently used features while keeping less-used options accessible for potential future needs.

---

Balancing Exploration and Usability

A key challenge in interface design is the balance between exploration and exploitation—providing reliable solutions while experimenting with potential improvements. For example, while payroll management might dominate a user's activity, interfaces should still offer visibility to lesser-used features like currency control or new services.

---

Conclusion: Towards a User-Centric Future

Intelligent interfaces can significantly enhance user experience, offering a level of personalization previously unattainable. For banks, this presents a complex but achievable goal. Overcoming legacy system constraints and fragmented data is critical. Once addressed, the potential of skilled analysts, developers, and data scientists can be unleashed to deliver user-focused, adaptive banking solutions.

By trusting in talent and embracing innovation, banks can transform client interactions, ensuring every user feels as though their banking platform is uniquely tailored to their needs.

Product list optimization project, part 1.3. GA tuning: Pagetype dimension

This post is a little diversion from the main course of this project.

Earlier I’ve been working on another UX research project which goal was to discover behavioral patterns in users’ sessions, in order to improve information architecture of the website. My approach was to apply hierarchical clustering to a number of sessions. But it’s hard to find any patterns when you parametrize session via URLs of each page in this session. Simply, it’s too much of diversity.

So I came up with idea to make a new dimension, Pagetype. Roughly speaking, it’s a place (coordinate) of this page in the website’s information architecture. It looks like this:

“<Senior section of the website> <level X> <particular type of the page>”

For example, “catalog level 4 product list”, “catalog level 5 product page”, “news level 2”, etc.

Performing some magic on distance measure and using Ward's minimum variance method while clustering, it gives very interesting outcomes for the information architecture analysis.

But let’s get back to our project!

I’ve mentioned Pagetype dimension because it helped here in this product list project too. Website I’ve been working on has product lists on different levels of catalog, from 2 to 5. So I used Pagetype custom dimension to filter product list pages from pages of other types.

This website happened to have a little bit weird CMS, and only javascript after page is loaded can decide whether it is a product list or not. So I’ve made a dedicated “Window loaded” tag in Google Tag Manager to track the Pagetype custom dimension.

I recommend to have such custom dimension if you work with behavioral patterns.

Learning Python for Data Analysis and Visualization course finished!

It finally happened! Over 100 lectures course Learning Python for Data Analysis and Visualization on Udemy. Very good introduction to Data Science with Python language. Lots of useful outbound links for the further exploration. And lots of stuff to think about, like in which tasks one should choose Python and where to choose R.

Kaggle Walmart competition

Just submitted my code on GitHub, which gave me the first Kaggle badge (top 25%).
https://github.com/Oleg-Davydov/kaggle_walmart_competition
I used Gradient Boosting with R package caret. But before that, there was a lot of preprocessing.
This was my second Kaggle competition. I finished 86th out of 485 participants, 1% of score difference with the winner.
Probably not bad for the start :)

Wine quality assessment app

Developing Data Products course, ninth and last in Coursera's Data Science Specialization, finished. As a course project, we had to build web-application with elements of machine learning. And then make a pitch presentation for that app. We were expected to use Shiny package and Shinyapps.io platform to publish our projects on the web.

I found an interesting dataset of Portugal wines' chemical characteristics and assessed quality in UCI repository. It gave me an idea that one can predict wine quality after chemical analysis.
Random Forest was chosen as a modelling tool. I hosted the model right on my github account:
https://github.com/Oleg-Davydov/winequality

And there you are! :)
https://laborant.shinyapps.io/winequality

Pitch presentation was restricted by five pages including the cover. I added a bit of humor in it :) Use right and left arrows on your keyboard to turn pages.
http://oleg-davydov.github.io/winequality/Rpresenter.html

Machines are learned!

It finally happened! Machine Learning course on Coursera finished. https://www.coursera.org/learn/machine-learning
Mixed feelings as after an excellent movie. It's great that the good guys have won, but it's very interesting what will happen next! :)

All lectures have been very engaging. I've watched many of them two or three times. Now I have a notebook full of lecture notes. Hope Andrew will eventually write a book which I will definitely buy.

All programming assignments were far from "annoying homework", but rather interesting little challenges and at the same time the continuation of lectures with many practical tips and tricks inside.

I'm proud to get maximum grade, 100%. As a course bonus, I've connected on LinkedIn with lots of my classmates from all around the world! My network has grown several times.

Want to say Thank You Very Much to our instructor, Andrew Ng, professor of Stanford University. That was an amazing experience!

Workout in a right way

Do you do physical exercises? Do you do it right? :)
Our course project in Practical Machine Learning course https://www.coursera.org/course/predmachlearn was about this topic. Several people put on an accelerometers and performed some exercises the right and the wrong way (consciously). So our task was to build the model to predict whether a person does barbell lifts correctly or incorrectly.
I used Random Forest and Gradient Boosting algorithms. RF has won in a Cross-Validation. Here is my report: http://oleg-davydov.github.io/PML_course_project/course_project.html

Traditionally, many thanks to our instructor Jeff Leek, professor of Johns Hopkins University!

Machine Learning

Enrolled in Machine Learning course on Coursera.
https://www.coursera.org/learn/machine-learning/
The instructor is Stanford Professor and Coursera founder Andrew Ng! Looking forward for a very fun couple of months :)