An artificial-intelligence (AI) bank leapfrogs the competition by organizing talent, technology, and ways of working around an AI-first vision for empowering customers with intelligent value propositions delivered through compelling journeys and experiences. Making this vision a reality requires capabilities in four areas: an engagement layer, decisioning layer, core technology layer, and platform operating model.
Previous articles in this series have explored the first two areas. The capabilities of the reimagined engagement layer1 enable the AI bank to deliver highly personalized seamless journeys across bank channels and within partner ecosystems. The capabilities of the AI-powered decisioning layer2 transform customer insights into messages and offers tailored to address a customer’s unique needs. The current article identifies capabilities needed in the third area, the core technology and data infrastructure of the modern capability stack.
Deploying AI capabilities across the organization requires a scalable, resilient, and adaptable set of core-technology components. When implemented successfully, this foundational layer can enable a bank to accelerate technology innovations, improve the quality and reliability of operations, reduce operating costs, and strengthen customer engagement.
We begin by summarizing the primary demands banking leaders should consider as they plan an enterprise-wide initiative to modernize core technology, data management, and the underlying infrastructure. Next, we examine the key transformations required to modernize the core technology and data infrastructure. We conclude by sharing 12 actions technology leaders should consider taking to ensure the transformation creates value for customers and the bank.
An AI-first model places demands on a bank’s core technology
Across industries, many organizations have struggled to keep pace with the demand for digitization, especially as consumers accelerated their adoption of digital channels for daily transactions during the COVID-19 crisis.3 Even before that, however, the financial-services industry has historically had mixed success in technology. Institutions that were early adopters and innovators in technology have built up a complex landscape of technical assets over decades and accumulated significant technical debt. Some institutions have tackled this challenge; many are behind the curve. Meanwhile, alongside the incumbents, an extremely active fintech industry has been constantly innovating and raising the bar.
Financial institutions that have shifted from being intensive consumers of technology to making AI and analytics a core capability are finding it easier to shift into the real-time and consumer-centric ecosystem. As AI technologies play an increasingly central role in creating value for banks and their customers, financial-services organizations need to reinvent themselves as technology-forward institutions, so they can deliver customized products and highly personalized services at scale in near real time.
At many institutions, standard practices now include omnichannel engagement, the use of APIs to support increased real-time information exchange across systems, and the use of big data analytics to improve credit underwriting, evaluate product usage, and prioritize opportunities for deepening relationships. As financial-services organizations continue to mature, the increasing demands on the technology infrastructure to support more complex use cases involving analytics and real-time insights are pushing firms to reexamine their overall technology function. Once they have committed to modernizing the core technology and data infrastructure underpinning the engagement and decision-making layers of the capability stack, banks should organize their transformation around six crucial demands: technology strategy, superior experiences, scalable data and analytics platforms, scalable hybrid infrastructure, configurable product processors, and cybersecurity strategy (Exhibit 1).
Robust strategy for building technology capabilities
Before embarking on a fundamental transformation of core technology and data infrastructure, financial-services organizations should craft a detailed strategy for building an AI-first value proposition. They should also develop a road map for the transformation, focusing on three dimensions of value creation: faster time to market with efficient governance and productivity tracking, clear alignment of demand and capacity to meet strategic and near-term priorities, and a well-defined mechanism to coordinate “change the bank” and “run the bank” initiatives according to their potential to generate value.
Faster time to market requires efficient and repeatable development and testing practices coupled with robust platforms and productivity-measurement tools. Aligning demand and capacity according to strategic priorities works on two levels. On one level, banks need to ensure that execution, infrastructure, and support capacity are optimized to ensure constant operation of all use cases and journeys. On the other, with constant uptime assured, work should be organized and scheduled to expedite projects having the greatest impact on value. Finally, financial institutions should establish clear mechanisms for setting priorities and ensuring that each use case is designed and built to generate a return exceeding capital investments and operating costs.
Superior omnichannel journeys and customer experiences
Building journeys that excite customers with their speed, intuitiveness, efficiency, and impact typically involves various applications spanning multiple bank and nonbank systems, all linked together by a series of APIs and integrations. This complex information exchange enables the organization to ingest valuable data from diverse sources to produce highly personalized messages and offers that speak directly to the customer in near real time. In addition to a standardized approach to managing APIs, banks should develop a clear mechanism to integrate across channels, core systems, and external interfaces while managing changes across multiple dependent systems. They should bear in mind, for example, that introducing a change in an existing digital channel could potentially entail changes not only across the front end but also across multiple interfacing systems, core product processors, and analytics layers.
A focus on journeys and user experience also benefits back-office and operations teams. New products are increasingly automated at the back end, freeing staff to focus on genuinely exceptional scenarios and differentiating activities, rather than repetitive low-value activities.
Finally, to ensure maximum value, use cases and capabilities should be designed as “enterprise products” to be reused in other areas. For example, the deployment of microservices handling discrete tasks like document collection and ID verification can ensure consistency in the way things are done across the organization. APIs should also be documented and catalogued for reuse. APIs that are domain- or product-centric (for example, enabling the retrieval of customer details from a single customer store) have higher reusability and take an enterprise-level view of the capability, as compared with a journey-centric API design—for example, one where an API supports retrieval of customer details for a specific mobile journey.
Modern, scalable platform for data and analytics
Delivering highly personalized offers in near real time requires AI-powered decision-making capabilities underpinned by robust data assets. What is more, the at-scale development of machine-learning (ML) models that are context aware in real time requires automated DevSecOps4 and machine-learning ops (MLOps) tools to enable secure and compliant continuous integration (CI) and continuous deployment (CD). This entails complex orchestration across source systems, data platforms, and data sciences to enable lab experimentation and factory production. This is particularly complex in a highly regulated environment where the involvement of security, audit, risk, and other functions is crucial in many stages of the process.
The incorporation of feedback loops with channel systems enables models to evaluate the output performance and make automated adjustments to increase the effectiveness of personalized messages, so the organization can generate personalized offers nearly instantaneously. For example, in the case of location-based offers for adjacent products, an organization must be able to overlay in real time customer location and preferences (as reflected in previous transactions) with predefined offers from nearby participating merchants.
Scalable hybrid infrastructure utilizing the cloud
With the continued expansion of customer engagement across bank and nonbank platforms, financial institutions need to create hyperscalable infrastructure to process high-volume transactions in milliseconds. This capability is made possible, in part, by infrastructure as code, automated server provisioning, and robust automated configuration management processes, which together solve the problem of “snowflake” configurations resulting from organic and complex linkages and changes that have accumulated over time.
Hosting these environments on a distributed-network cloud environment allows a balance between paid-up-front baseline storage and computing capacity, on the one hand, and, on the other, elastic on-demand surge capacity without disruptions to service. Self-monitoring and preventive maintenance also are automated, and disaster recovery and resiliency measures run in the background to ensure constant uptime even if incidents evade automated self-repair and require manual intervention. As a result, the risk of disruption to critical operations is minimized, and customer-facing applications run with high availability and responsiveness. The combination of on-premises and cloud-based infrastructure is increasingly relevant in high-volume and high-frequency areas such as payments processing, core banking platforms, and customer onboarding systems. Making workloads “cloud native” and portable allows the work to be moved to the most appropriate platform.
Highly configurable and scalable core product processors
To sustain a leading-edge value proposition founded upon AI and ML capabilities, banks must continually evaluate their core products and identify opportunities for innovations and customizations. Combined with deep understanding of customer needs, enabled by advanced analytics, an organization can anticipate emerging customer requests and design distinctive products accordingly. The need for real-time reconciliation and round-the-clock transaction processing also emerges as a key competitive advantage for financial institutions. For example, with the advent of next-generation core banking platforms, organizations can now develop products that are built for scale and can be readily configured to meet specific customer expectations.5
Secure and robust perimeter for access
It is crucial to ensure that the organization maintains an appropriate cybersecurity posture across the entire technology infrastructure as protection against vulnerabilities within applications, operating systems, hardware, and networks. Financial institutions should also implement appropriate measures to secure the perimeter and control access to various systems and applications within the organization’s infrastructure footprint, including private and public cloud servers and on-premises data centers. For example, transferring workloads from traditional on-premises infrastructure to public cloud requires careful measures to protect customer data, along with a robust strategy for detecting and remediating potential threats and vulnerabilities.
The “classical” approaches of securing the perimeter should be coupled with more modern approaches to limit the impact of intrusions or reduce the “blast radius.” Again, AI has a part to play here, given the advent of increasingly sophisticated network intrusion detection, anomaly detection, and even forensics during postmortems of security incidents.
Start the transformation by prioritizing key changes
To meet these demands, financial institutions will need to transition from a legacy architecture and operating model to an automation and cloud-first strategy. Building the core technology and data capabilities upon a highly automated, hybrid-cloud infrastructure can enable the AI bank to scale rapidly and efficiently as it gains competitive and differentiating capabilities.
The AI-bank capability stack combines core systems and AI-and-analytics capabilities in a unified architecture designed for maximal automation, security, and scalability. Getting to this target state requires a series of complex initiatives to transform the organization’s core technology and data infrastructure. These initiatives focus on several key areas: tech-forward strategy, modern API and streaming architecture, core processors and systems, data management, intelligent infrastructure, and cybersecurity and control tower (Exhibit 2).
Banks should begin this far-reaching initiative by translating the AI-first vision into an enterprise strategy that merges technology with business, funding investments in innovation with the returns on incremental changes in technology.6 Business and technology collaborate as co-owners in designing and managing operating models and outcomes. This “tech-forward” mindset thrives in interdisciplinary teams focused on innovation and led by skilled engineering talent leveraging modern tools and practices for first-time-right releases. Organizations should also adopt enterprise agile practices for high-velocity engineering teams, with integrated cross-functional teams of business, technology, and functional experts, and external partners using modern approaches to software development, testing, release, and support cycles. In addition, efficient management of the full stack requires governance of the technology function through a standardized set of metrics, along with ongoing tracking of uptime and health for each component of the stack.
Modern API and streaming architecture
Next, banks should integrate internal and external systems to support seamless customer journeys across internal platforms, partner ecosystems, and numerous external interfaces. This requires a robust, scalable, and standardized approach to building and hosting integrations and APIs. The APIs, in turn, should be rigorously tested for performance and developed using agile release principles. When a well-defined stock of APIs-as-products are orchestrating flows across systems, product innovations can advance from concept to production and deployment of minimum viable product within 30 to 60 days.
To complement a robust API strategy, technology leaders should also consider establishing a high-speed data-streaming channel to enable standardized asynchronous data transfer across the enterprise in real time.
Core processors and systems
With the right architecture in place, banks can shift away from traditional, complex, and tightly intertwined core systems to lightweight and highly configurable core product processors and workflows. These processors are also complemented by “microservices,” or discrete applications (such as for payments, card accounts, or loans) that “externalize” the logic within traditional core platforms.
The transition to lightweight core processors and systems hosted on scalable, modular, and lean platforms exposed as APIs supports, for example, real-time reconciliation and allows changes to be made in live systems with zero downtime. Use of modern cloud-based infrastructure to host such platforms also makes it easier to scale up. If successfully implemented, a lightweight processor platform can enable an organization to advance from new-product concept to launch in two to three months. This is a significant advantage against organizations constrained by legacy technology, where launching a new product or customizing an existing product can take six months or more. Assembly of new off-the-shelf product stacks can also enable innovative new customer propositions, such as an end-to-end lending journey on a modern stack using these principles.
Data management for the AI world
It is crucial to establish a modern data and analytics platform to fuel the real-time ML models of the decision-making layer. The analytical insights generated by these models are deployed through martech tools to craft the intelligent offers and smart experiences that set an AI bank apart from traditional incumbents. In order to support superior omnichannel customer journeys and seamless integration with partner ecosystems, the data platform must be capable of ingesting, analyzing, and deploying vast amounts of data in near real time.
The data platform should also provide scalable workbenches with AI and data-science capabilities to lab and factory teams. These workbenches enable teams to access relevant data sets as they develop models and deploy insights in product iterations. The infrastructure should also support the development of ML models through automated and repeatable processes.
If an organization allows interdisciplinary teams across the enterprise to search and extract data held on the platform, these teams can optimize their data consumption according to customer needs and market opportunities. It is essential to enable data-science teams with appropriate tooling and access to scalable computing power so that they may experiment and innovate. Underpinning these actions, appropriate technical documentation and cataloging of assets (for example, APIs, ML models, data dictionary, DevOps and MLOps tools) ensure proper governance and access control. By creating ML models and scorecards through a well-defined lab-factory model, AI-first organizations empower employees to leverage self-serve, real-time data and analytics infrastructure to guide value-based planning and support daily decision making.
Banks then should ensure they have an effective strategy to modernize infrastructure. For this, they should consider the adoption of public cloud to complement the traditional infrastructure in situations where workloads require resiliency, scale, and use of hosted or managed offerings (such as hosted databases). Public cloud enables velocity through higher levels of automation, templates, and reduction of operational risk. When setting up such environments, banks must build upon the foundational elements of infrastructure management, including observability, resiliency, and high availability, as well as a robust configuration strategy. A well-tuned, scalable, and load-balanced stack can support response times of less than a second while scaling horizontally to cater to variations in transaction volume.
Cybersecurity and control tower
Finally, institutions should address cybersecurity and control. This includes setting up a centralized control tower to monitor data, systems, and networks across the infrastructure. The scope of responsibility includes ensuring boundary security and identifying and rectifying threats and intrusions. Also crucial is to establish a well-defined set of compliance measures for security testing and vulnerability scanning before deploying assets on live systems. These measures reduce the risk posed by potential threat scenarios.
Technology leaders should prioritize interconnected capabilities
Given the broad scope of components to be transformed, organizations should bear in mind that optimal outcomes are much likelier when they first establish a holistic strategy for technology transformation. Unfortunately, not all have found the resources to embrace fully the potential offered by the rapid advancement of AI technologies and the steady rise in customer expectations. Some financial institutions, despite seeing the imperative to change, have maintained and modernized their legacy platforms. Various business lines have set up organically built platforms upon this foundation, making it costlier and more and more complex to maintain. Many organizations have spent billions of dollars on multiyear technology initiatives within silos, only to find that they fail to generate the scale benefits required to justify investments. Leaders should heed these lessons, adopt a holistic perspective, and map priorities according to the end-to-end impact that each step in the technology transformation has on the value of the enterprise.
If an organization meets the strategic demands outlined at the top of this article, the implementation of modern core technology and data infrastructure can yield significant value in the form of faster delivery of changes and improvements, increased cost efficiency, higher quality of assets, and stronger customer outcomes. For example, a sound DevOps and release-management strategy can contribute to a 25 to 30 percent increase in capacity creation, a reduction in time to market of 50 to 75 percent, and more than a 50 percent reduction in failure rates.7 In turn, development efforts can improve schedule adherence by 1.5 times and reduce customer defects by 20 to 30 percent through process automation and agile ways of working,8 and leading organizations have improved issue-resolution time and planning time by between 30 and 50 percent.9 There are indirect benefits as well: by empowering employees with a clear mission, autonomy, and strong focus on customers, agile organizations have been able to increase employee engagement by 20 to 30 percent, as reflected both in willingness to recommend their workplaces and in employee-satisfaction surveys.10
Technology transformations are fraught with risk, including delays and cost overruns, and only those organizations whose leaders are prepared to commit the energy and capital necessary to carry through with the comprehensive effort should embark on the journey. Ultimately, this is a decision not just to survive, but to thrive, and it requires a change in mindset. Specifically, traditional financial institutions will need to break out of their legacy technology architecture and explore AI-and-analytics opportunities. Should they undertake the challenge and begin thinking about how best to chart their course to becoming an AI bank, their leaders may consider 12 key insights gleaned from the experience of financial-services leaders that are in the process of carrying out such transformations (Exhibit 3):
- Consider the factory model to build at scale. Leverage a factory approach in fast-evolving and critical areas of the transformation to enable repeatable execution and development of capabilities within technology teams and to promote standardization to speed up execution. For example, a core system factory consisting of teams, predefined operating procedures, and systems to manage, prioritize, and execute changes across business units can expedite deployment of new solutions significantly.
- Consider insourcing differentiating capabilities. Based on the eventual outcomes desired, build certain differentiating capabilities in-house, with robust engineering support, perhaps starting with APIs, infrastructure, or the data and analytics platform.
- Maintain rigorous documentation on integrations. Remember that the development of engagement systems and comprehensive changes in core-technology require significant adjustments to integrations, and substandard documentation of the specifications for these integrations often slows the broader initiative to transform the bank.
- Identify an anchor stack but experiment with others. Emphasize the importance of standardization for engineering-centric development at scale, and build on a single stack to support faster change. At the same time, continue experimenting with other stacks and stack components for smaller builds in order to adopt alternative or newer approaches where the incremental benefits are clearly defined.
- Maintain an automation-first and fast-release posture. Adopt an automation-first and frequent-deployments posture on fast-evolving applications and stacks. While initial hiccups are not uncommon, release rails should be hardened over time to speed up time to market. Well-defined release management and deployments are key to execution velocity. Standardizing through DevSecOps typically unlocks productivity gains of as much as 20 to 30 percent.
- Consider a modern core for high-velocity areas. Consider modern and lightweight core systems built on scalable and hybrid infrastructure to enable an efficient rollout of new capabilities while enabling a modular build of financial products.
- Adopt a value-centric approach to building data platforms. Take advantage of the fact that data and analytics platforms evolve over time, and do not allow teams to be overwhelmed by the rapid shift of tooling and available technology. We have observed that organizations that budget the anticipated return of change efforts are able to prioritize use cases that are functionally simple, fit the road map for building the platform in iterations, and realize economic value along the way.
- Set up a lab and factory for analytics. Establish a lab to experiment with tools and platforms for efficient development in test-and-learn cycles. Also, build a central factory for producing and deploying analytics use cases at scale on an individual stack.
- Define the enterprise cloud strategy. Create a common strategy across stakeholders to enable a structured and systematic migration to the cloud. Cloud adoption poses multiple firsts in the enterprise in terms of security perimeters, change management, and cloud-migration and disposition strategy.
- Establish end-to-end visibility across the technology and infrastructure stack. Recognizing that at-scale digital transformations impose limitations on volume and scale, implement robust automated tools to observe stack performance and to diagnose and resolve issues.
- Identify the right perimeter design for the cloud. To safeguard against potential malicious attacks on cloud-based public-facing applications, design an appropriate network perimeter that optimizes the potential attack radius.
- Ensure data security on the cloud. Design robust data-categorization and data-security safeguards to avoid critical customer-data combinations and comply with national data-protection and data-residency laws.
If banks are to thrive in a world where customer expectations are increasingly shaped by the AI-and-analytics capabilities of technology leaders, they must rebuild their core technology and data infrastructure to support AI-powered decision making and reimagined customer engagement. These are the three “technology layers” of the AI-bank capability stack. The full stack also includes a leading-edge operating model to ensure that all layers work together in unison to deliver intelligent propositions through smart servicing and experiences. The AI bank of the future requires an agile culture and platform-oriented operating model that respond promptly to emerging opportunities and deliver innovative solutions rapidly at scale. The next article in this series examines the crucial elements of the platform operating model.