In the face of digital transformation, banks have significantly transformed their business models, culture, and operational models – taking them beyond the traditional banking systems. Faced by ever-changing regulations and a rapidly evolving technology landscape, banks are breaking down the legacy-era monoliths and adopting microservices as the core banking platform.

The digital world brings with it the complexity of handling large amounts of data generated from millions of users. Despite the rise in real-time analysis and superior computational algorithms, traditional data architecture cannot capture all enterprise and external data. To complement existing data architecture services, banks require event-driven rules to unite functionalities.

Traditionally, banks have followed a request-driven model wherein a rigid architecture defines tasks. These systems are efficient in developing simple and set tasks, but fail to react to variable cases of the digital era. For example, a user accesses your banking portal for fund transfer. During their time on the portal, they might get interested in a different product. They seek out the product, read more about it, but eventually forget about it/lose interest and move on. The traditional request-driven architecture would fail to identify this opportunity of business interaction; leading to loss of potential sale.

In order to prevent such loss of data, developers are moving to an event-driven architecture (EDA) system. But what is EDA?

Gartner defines EDA as “a design paradigm in which a software component executes in response to receiving one or more event notifications. EDA is more loosely coupled than the client/server paradigm because the component that sends the notification doesn’t know the identity of the receiving components at the time of compiling.”

An ‘event’ is a notable thing that can occur inside or outside of a system, triggering a set of services, business processes or operations, while event processing deals with detecting and responding to events that have meaningful business outcomes. Taking up the previous example – With the EDA setup, your customers’ interest in products gets registered as an event by the system. Based on the event captured, banks can now categorize customers into prospective clients and open new lines of business interaction for your sales or business development teams.

Rabobank, a Dutch multinational banking and financial services company, has been continuously working on its real-time event-driven bank. Moving away from the traditional batch processing, Rabobank is building a Business Event Bus (BEB) to share business events across the organization. Effective communication with its millions of customers was a scalability and flexibility issue that the bank was able to overcome by adopting EDA and event programming into their mainframe. The bank developed ‘Rabo Alerts’ – a system to alert customers in real-time whenever an interesting financial event occurs and thereby, drastically reducing customer alert timing from 5 minutes to just a few seconds.

In the past few years, EDA has increased in popularity owing to changing markets, connected consumers, and mobility. The architecture is used to build reactive applications that are event-driven, scalable, resilient, reliable, distributed, and interactive. The real-time application communicates asynchronously across systems wherein the sender and receiver, both, remain anonymous. Since systems are triggered only in case of an event, EDA enables loose-coupling between components; eliminating dependency and lowering operational costs.

While consumer service is a push for EDA, regulatory compliance is also becoming a data-driven discipline. Banks and financial institutions continue to struggle to comply with regulations such as Bank Secrecy Act/ Anti-Money Laundering (BSA/AML). The high costs of non-compliance is not a factor for consideration for any agency. Issues from lack of uniformity in the KYC process to lack of periodic assessment of vendors are all areas of potential non-compliance. But at the core of the system lies KYC.

After onboarding a customer, the KYC system needs to perform regular on-going monitoring, periodic customer risk re-assessment, and re-certification. For an event-driven enterprise, this does not pose a challenge as an EDA would take care of submitting events to a monitoring platform. The system will be able to filter events of interest and submit them to respective business units – setting off a cascade of business processes to reassess customer’s risk and re-certify the customer.

By becoming event-driven, enterprises specifically banks improve their scalability and regulatory compliance. The evolution of EDA to integrate microservices or API management solutions adds another dimension to the development of a smart Service-Oriented Architecture (SOA). Increasing adoption of IoT and big data will further boost EDA in the banking sector.

The recent introduction of PSD2 (Revised Payment Service Directive) has banks navigating the unchartered territories of open banking. A PWC report predicts 71 percent of SMEs and 64 percent of adults will adopt open banking by 2022. The open API economy fosters the development and delivery of products and services through collaboration with third-party entities.

To stay ahead of their competition, banks are adopting the open API model to re-architecture existing applications along their IT transformation journey. The beauty of microservice architecture lies in the freedom of decision making in quickly deploying independent service entities. Some banks such as Fidor and Atom & Starling, are adopting an API-first approach to product development. Their Banking as a Platform (BaaP) redefined traditional banking by leveraging APIs and microservice architecture. The API-first paradigm requires the adoption of technology-agnostic design practice. The approach does not rely on any programming language, technology platforms (SOAP, CRUD, REST etc.) or libraries.

At this juncture, adopting a design thinking approach is helpful to identify and create interactive systems focusing on the users, their needs, and their requirements. Implementing a design thinking approach to microservices architecture is where human-centered design meets development. Consumers are able to interact with baking applications in early stages providing valuable feedback for product development.

The concept of design thinking is relatively new to the banking sector; its implementation brings in new business and cultural challenges that the industry is not used to. Some organizations develop their own microservice design canvas to guide service development. The canvas serves as a tool to help capture critical attributes of a service before the development phase. Taking an outside-in approach, the schema consolidates key insights on consumer tasks, interface requirements, non-functional qualities (security, availability, reliability, scalability etc.), processing logic, data elements, and external service dependencies into its design.

Some banks, like Capital One, have adopted the open banking systems, launching a developer portal (DevExchange) and three open APIs – SwiftID, Rewards, and Credit Offers. SwiftID operates as an easy to integrate two-factor authentication security system, whereas the other API services provide information pertaining to rewards and credit card offers. The idea behind creating the commercial banking platform was to create a tool for next-generation CFOs to simultaneously access, modify and analyze information across channels. Using the microservice design canvas, Capital One was able to align its goals and assimilate design thinking early into their development strategy. The tool helped in dealing with the complexity of distributed software and understanding the evolving microservice boundaries.

But moving from a monolith service to the discrete systems of microservice architecture is not easy. We explore the prominent challenges of design thinking in microservice architecture.

Breaking down the monolith

Banks have been maintaining and operating monolith applications or legacy systems for a long time. Bringing in the microservices architecture is a challenge as it requires an investment of time, money, and a shift in mindset. Compared to the monolith architecture, a microservice architecture breaks down large software projects into inter-communicating modular units through APIs.

Decomposition of a monolithic application can be complex. The first question being on which service should get broken down to an individual service unit. One approach would be to partition the service requirements along business functionality lines. For example, if you have an investment lookup function with multiple calls from other functions, it’s a starting point for breaking it out into its own service.

Thinking like an engineer

Most developers begin building services from an inside-out approach; designing the data layers first and leaving the API design for the end. This approach takes into account the business logic and services it offers but does not look into a user’s needs. The engineering-thinking approach follows a set path of logic for design. Whereas, the design thinking process is a systematic approach to enable creative thinking and solution development. It involves the interaction of diverse personalities and talents towards creating the best solution for the user.

The design thinking process highlights the central role of users/customers and how new products can solve their needs. Using an outside-in design involves real users from the beginning of development to testing various scenarios.

Poor definition of the problem statement

A key part of the design thinking process is defining a clear and focused problem statement. Your problem definition paves the way for ideating and designing APIs. Once an API is released, customers and business units are dependent on its functionality. Changing API functioning after its release cannot be done without disrupting systems. Developers, on the other hand, get stuck in the integration and testing loop, giving them little time to develop new applications.

Due to this, the first parameter of a microservice architecture is to define its functionality – ‘what it should do’, ‘how should the services be split’ etc. Organizations have to decide which microservice approach works for them and supplement it with the correct infrastructure. Some organizations prefer a monolithic core model whereas others prefer a fully-service oriented approach.

Testing in a microservices environment

A business application is made up of multiple microservices interacting with each other.  Apart from testing the overall application, developers are required to test each microservice separately. Additional layers of testing and API integrations creates a complex testing process; requiring more quality assurance (QA) engineers and extensive planning. The interdependencies of a microservice make it difficult to ensure full coverage of test cases, especially in lower environments. Indistinct behaviors are harder to predict, validate and monitor on a regular basis.

Organizations have to work on developing a comprehensive approach to their microservice testing environment. Unit tests and integration test take care of coding and integration points of a microservice architecture. They ensure that every piece of code, down to the lowest levels, and entire objects are tested before integrating with other parts of the application.

Conclusion

In the face of changing customer behavior, regulations, and technology, banks are on the lookout for models that will be the ‘Uber of Banking’ for their challenges. Microservices make this possible by their human-centric design API designs. The approach supports the architecture with an agile, scalable, and dynamic service mesh. In this atmosphere, design thinking is proving to be a useful tool to help banks recognize the needs of their users and other stakeholders. For banks willing to experiment with design thinking, partnering with external design support teams and microservice developers can be a start.

In today’s world, banks are continuously competing with startups and Fintech disruptors. The rapid growth of Fintech start-ups and new incumbent financial institutions poses a threat to the traditional banking operation model. To keep up with the demands of the digital age, banking requires a new way of thinking to address customer needs and challenges.

Although the banking industry is not new to customer-centricity, there lies a difference of how the sector approaches its solutions. Compared to the customer-first designs of Fintech, traditional banking design of products begins tech adoption with an inside-out approach of resolving internal operational efficiencies and then extending the same to the customer.

The banking industry is moving toward open banking wherein, application programming interfaces (APIs) are used to share consumer financial data to third-parties. At its core, APIs are fundamental to the concept of open banking. As user-interfaces are important to consumers, the design of APIs greatly impact how developers interact with it. A well-designed API would assist developers to work efficiently, provide a quick turnaround to develop solutions, and foster innovative product develop for consumers and fellow developers.

Design thinking is a customer-centric approach to the identification, design, development, and evolution of products and solutions. Earlier business challenges where approached in a linear fashion, you analyze a situation and decide. The non-linear approach of design thinking involves a defined process and iterative understanding of customer needs. Teams now iteratively design, ideate, prototype, evaluate, and decide. The discovery process is deeply integrated with the organization’s innovation process and continuous feedback from the customers.

Typically, design thinking follows four or five-stages depending on which school of thought one follows. The IDEO process uses a four-phase model of question, care, connect, and commit. Whereas, the Hasso-Plattner Institute of Design at Stanford (d.school) proposes a five-stage model: Empathize, Define, Ideate, Prototype, and Test.

Empathize

As banks start to develop a deep understanding of customers’ needs new paths to innovative services, empathy plays a key role. In this stage, organizations identify and engage with their customer segments to understand their pain points, motivations, and needs. Conducting research based on user personas would also help in better connecting with the customer. For example, banks can look at assessing how their customers prefer to communicate

Define

Insights from the empathize stage provide clear and focused problem statements that need to be addressed. Upon analyzing the information, organizations can define the core problems faced by the customers. Based on user feedback, financial institutions can highlight areas of improvement and define clear goals for their teams. Instead of asking “We need to increase customer-chatbot interactions?”, banks can define the problem such as “Our chatbots technology should factor in elements of semantic-oriented systems”.

Ideate

A definitive problem statement gets the ball rolling towards ideation phase. At this stage, teams brainstorm on ideas geared toward human-centric solutions. New solutions are discussed, and the problem is assessed from all angles. A wealth of ideas can be amassed during the ideation phase from which a few can be prototyped and tested.

Prototype

Design teams can now work on developing an inexpensive minimally viable product (MVP) based on the solutions suggested in the previous stage. In this experimental phase, the prototype can be tested across teams, departments and external validators to identify the best possible solution. The prototype stage gives a hands-on experience for teams to interact with the end product, identifying issues and going back to the drawing board to investigate further.

Test

In the testing stage, organizations introduce their solutions to end-users. The iterative process involves the customers at an earlier stage of product development, working closely with the design team. Based on the results of this stage, banks can now evaluate or redefine problems to further enhance the product for the customer.

The five-stages of design thinking serve more as a guideline for organizations to adopt the concept. Banks have started to recognize design thinking as a core tool for innovation. Leading banks are hiring design teams to transform their workforce, products, and services. For example, Spanish banking giant BBVA has over 1000 staff ambassadors to promote design thinking practices across its organization. Whereas, banks such as USAA have invested in setting up 120-person design studios to develop its design community.

Design thinking is not just a company policy; it is a cultural mindset. Design thinking facilitates rapid innovation through a fine balance of structure and creativity for problem-solving. For successful payoff, the concept of design thinking needs to inculcate across the organization. For example, Citi is working with design consult IDEO to train its employee on design thinking with agile methods for innovation. With design thinking, banks are able to improve customer relationships along with its value proposition.

Over the course of time design thinking has evolved from being a scientific approach to design to a human-centered service design. Currently, design thinking is being applied to solve complex business problems; serving as a differentiator in competitive landscapes. The future of design thinking would see its application across complex environment of human behavior and push boundaries of customer experience.

Today, as digitization sweeps through the banking industry, CXOs are working on digital strategies to accelerate business performance and operations. While customer-centricity has been the mantra of the banking industry, evolving technology and customer needs pose a problem establishing banking services and products.

With digital banking and omni-channel experiences becoming common expectations, banks have to effectively strategize their digital resources. But tackling the digital landscape can be daunting. A recent survey by BCG stated that only 43% of organizations have a “clear digital strategy for the corporate bank as well as a well-defined roadmap for digitization”.

To decide the best digital transformation strategy for your organization, PwC recommends three approaches:

• Front end only – CX improvements with no changes to the IT infrastructure
• Wrap and digitize – Digitization of individual components of banking functions
• Go digital native – Digital customer interface with a complete digital back end

Front-end only

Today’s tech-savvy customers are used to seamless online transactions and one-click applications. Fintech see a growing customer base with their offering of a wide range of services with an easy to navigate interface. For banks to compete in this environment, the first simple approach would be to digitizing the customer interaction channels. These include building a website or a mobile app to cover basic aspects of customer interactions with the bank.

The focus of this approach is on only improving the  CX and customer-facing systems. Viewed as a ‘cosmetic-fix’, helps banks put up a digital front without having to invest in changing the legacy systems. It is one of the quickest approaches and would give the banks the initial push towards digital transformation.

The cosmetic digital front has helped banks stay afloat in a competitive environment. But for true success, digitization of front end should be swiftly followed by integration of back-end systems to significantly improve customer service. Back-end operations cannot hide behind the digital front for too long without increasing operational costs. In order to scale up, back-end systems have to be updated.

For example, Digibank introduced the first mobile-only bank in India. The application implemented biometrics and artificial intelligence (AI) for a seamless paperless, signature-less and branchless bank. The easy onboarding process of less than 90 seconds has helped increase their customer base to over one million in just one year. The success of Digibank is a mix of right technology and marketing. Currently, Digibank offers limited products and services such as money transfers, bill payments, and ATM services. But for the organization to become a standalone digital bank would depend on its back-end process. As the bank looks to further expand its digital transformation efforts, it is working on alternate digital strategies to solve its sustainability issues.

Wrap and Digitize

A longer process to digitization would be the Wrap and Digitize approach in which, the front-end digitization is supplemented with replacing/upgrading legacy infrastructure with digital technology. This approach is effective as it integrates the middle and back offices during the setup process.

Banks are adopting the ‘Wrap and Digitize’ approach as it significantly improves customer experience compared to the front-end only approach. The process focusses on improving individual components of banking functions. As per PwC, banks can use APIs to integrate their data, functions, services, and products under one roof. The flexibility provided by the API enables the organization to be more agile in its operation. For example, the next phase of digital transformation for Digibank involved improvements in their back-end systems. A new code was built over the existing back-end assets to support rapid growth and expansion strategies. The new architecture enabled API-based banking and a digital platform for banking through multiple business partnerships.

The lengthy process of wrap and digitize can be a deterrent for banks. Every process is addressed one-by-one before moving onto the next one. The transformation process can take several years to be completed. Despite the long duration, the process is the most cost-effective as the investments are spread out over a period of time. The gradual approach of integration works as the best option for traditional banks and credit unions.

Go Digital Native

PwC’s digital native approach is for challenger banks and digital-only banks poised for rapid growth. Banks start by creating minimally viable banks (MVBs) that offer limited services or products. By being digitally native, banks are built on a digital core and open architecture enabling the development of a fully-functional agile organization. This approach emphasizes on customer-centricity and enables banks to shift their operations based on customer preferences.

The functioning of a digital bank would require an infusion of digital mindset into the traditional banking atmosphere. Digital banks also face significant regulatory issues that have been written for traditional banking sectors. For example, in the United States, digital banks are expected to meet the same regulatory standards, reporting and consumer protection regulations as incumbents. But agile organizations like DBS, are using third-party services that leverage AI natural language processing (NLP) to manage regulatory compliance across governing bodies.

Digital disruption in banking is pushing banks to rethink its business models. To follow the path of digital transformation banks would have to determine their long-term strategy before choosing the technology. While there are several ways to decide the path to digital transformation, we outline a basic guideline for organizations:

• Multidisciplinary thinking – To ensure the viability of an organization, develop a long-term strategic plan taking different perspectives of business, IT, compliance, and operations into consideration. The multi-faceted approach enables the development of an agile organization that is able to support quicker product development. Adopting a digital-first strategy enables growth that is not limited to existing technology and channels.
• Functionality list – Deciding on the best digital banking platform is easier when the functionality of each platform is listed and understood. The solution selected should align with your long-term strategy, requirements, and goals.
• Vendor cultural fit – Analyse your strategic partnerships and vendors for cultural fit with your organization. Vendors that leverage innovative technology are preferred as they enable an open-architecture approach. Connecting the ecosystem to a third-party service future-proofs your bank from modern technology disruptions.
• Involve internal stakeholders – Get your teams to work closely with internal stakeholders, like senior management, to develop minimum viable products in the implementation phase itself. This approach ensures the reduction of time and cost while gaining valuable insights through customer testing.

Digitization is radically changing traditional banks and credit unions. Many banks have recognized the importance of differentiated digital strategies and working toward complete digital transformation.

Artificial intelligence (AI) and Machine Learning (ML) technologies are rapidly transforming business functions, including software development. In its search for efficiency, the industry has slowly shifted from the traditional Software Development Life Cycle (SDLC) to an agile development environment.

Over the last decade, DevOps has become an industry standard. Its main goal has been to improve product delivery/development by encouraging communication between software developers and IT operations. Across industries, the need for developers-operations balance has led to the rise of DevOps ethos. According to Grand View Research, the DevOps market size is expected to reach $12.85 billion by 2025 with an 18.60% CAGR.

For a culture focused on efficiency and automation of tasks, it’s no surprise that AI and ML find their application in DevOps. From enhancing continuous feedback loops to software testing, AI/ML and DevOps complement each other perfectly. As IT operations become more agile and dynamic, AI can iron out the kinks in the system.

Gathering Key Data Insights

A DevOps process generates significant amount of data across servers and logs. Combing through big data to find specific instances can be cumbersome and not cost-effective. In this data deluge, ML optimizes application environments stay afloat with real-time data analysis.

Using supervised learning and training data, developers would be able to identify errors that would otherwise be missed in large data clumps. Machine learning can also be employed to analyze insensible data to identify patterns and behaviors that can be used towards data analytics. As ML reduces noise-to-signal ratio data silos are broken down for teams to use across product development. Developers are no longer limited by self-defined thresholds, giving room to assess data trends.

Correlation across Monitoring Tools

As dev teams expand, multiple monitoring tools are used to assess data and also check application health and performance. The layered algorithms of AI/ML accept multiple data streams allowing correlation of data across multiple monitoring tools. ML systems connect disparate data systems to provide real-time health assessments of applications.

Optimizing DevOps process

Employing adaptive-ML, DevOps teams can optimize specific values or metrics towards a certain goal. Neural networks are trained to maximize a single value/ parameter; enabling the system to adapt and change during the production phase itself. This ensures the optimization of values throughout the development lifecycle.

Enhancing Security and IT operations

Mining of large complex datasets helps in gaining meaningful insights in predicting product and server failures, avoid technical drawbacks and facilitate decision making in DevOps business framework. Most security protocols are implemented the end stage of the development life cycle. In case of banking sectors, DevSecOps – the culture of integrating security within the DevOps process, is gaining ground.. This philosophy emphasizes on ‘security as code’, allowing streamlining testing parallelly to security and compliance reviews. The use of AI-based digital security technology allows banks to meet market demands while continuously monitoring potential security risks.

Smarter Resource Management

By automating routine and manual tasks, AI/ML systems aide in efficient resource management. Teams have more time to concentrate on efficient development and coding practices.

Software Testing and Shift-Left

Software testing is another aspect wherein AI/ML applications can be leveraged to evaluate coding errors in test results. Test automation is a critical part of shift-left testing. AI helps in running multiple diagnostic tests to identify the cause of failure. In the case of a test failure, AI is in a position to rectify the error even before the product hits the market. This approach involves minimum human intervention. Whereas, the continuous feedback loop aides developers write efficient code, reducing errors thrown by the system.

The capabilities of AI/ML-driven solutions in conjunction with DevOps is ever expanding. As organizations work towards identifying bottlenecks, hiring skilled talent is the impeding factor for the AI-driven organizations. On the foundation of a strong DevOps infrastructure, AI/ML will be synergic tools to increase efficiency. AI will continue to make inroads into more business cases with vertical-specific solutions that would transform business processes.