Looking beyond desktop integration: low-code, no-code, and continual learning

Financial services firms need to utilise code to develop a technology platform which bridges modern and legacy applications on the desktop

by | April 20, 2022 | Glue42

Desktop technology has progressed significantly over the last decade, thanks to ever-more-powerful CPU and GPU technologies (from innovators such as Intel, AMD, Apple, and NVIDIA) and the emergence of high-performance JavaScript engines powering functionally rich Progressive Web Applications.

Meanwhile, desktop-native application technologies have also continued to advance, thanks to their ability to function offline, take better advantage of hardware resources, and most importantly, take full advantage of native operating system capabilities.

This somewhat inevitable dichotomy calls for a technology platform which seamlessly bridges modern and legacy applications on the desktop. An increasingly complex business domain, driven by complex instruments being traded, increased data volumes, enhanced regulatory regimes and surveillance are but some of the factors driving significantly increased cognitive burden on financial services professionals. The traditional lack of interoperability between applications leads to constant application and context switching, cut-and-paste boondoggling, and needless user error.

Glue42’s market-leading Desktop Integration Platform solves this problem with a powerful framework that enables inter-application communication, workspace management, alerting and notifications, and a host of other capabilities required to enable disparate applications to coexist. Conformant to the FINOS/FDC3 Open Interoperability Standards,  Glue42’s platform enables web-based, native, legacy, and modern applications to seamlessly integrate and deliver end-user and domain-oriented business process workflows that promote a focus on high-value tasks, increase user productivity and reduce risk.

“Every desktop needs application integration,” says James Wooster, chief operating officer of Glue42. “The problem is that we have got so used to swapping, copying, and pasting our way between applications that we have become oblivious to the issues that this causes.”

Vendors need to recognise  the average trader will be using more than 10 apps during the day, with another 20 or 30 accessible for specific scenarios, says Wooster. Therefore, the ability to work seamlessly with other systems is critical.  “This should have far-reaching consequences over time as end-users begin to demand more consistent UI design and style to help navigate within and between these apps,” he says.

The low-code/no-code revolution

Recent years have seen the emergence of low-code and no-code application development methodologies. These approaches advocate minimal to no custom coding in the development of functionally rich end-user applications.

This is achieved using visual interfaces and widgets that implement rich, discrete units of functionality that can be drag-and-drop integrated into visually modelled business-centric workflows. The promise of these approaches is reduced time-to-market, increased automation, and ease of testing, and ultimately, lower total cost of ownership.

Low-code and no-code approaches can be particularly well-suited to the Financial Services domain, where cohesive, functionally complete, and efficient workflows are of paramount importance. By developing a Lego-like toolbox of widgets, which implement key functionality such as Options Pricing (computational), Volatility Surfaces (visual), Research (content-oriented), etc, firms can position themselves to deliver domain-specific applications at an accelerated pace and with better quality.

Glue42’s Desktop Integration Platform is a natural enabler of low-code and no-code, owing to its rich capabilities in workspace management, visual development, and interoperability driven by strict adherence to the FINOS/FDC3 standards. Several Glue42 enterprise customers have successfully adopted the platform to deliver integrated application suites to their end-users and are poised to expand this foundation into the next generation of low-code/no-code applications.

Desktop interoperability is critical to the success of low-code/no-code efforts because of the inherently disparate nature of the trader’s desktop. “Unless you are one of the few dominant software vendors in the financial services industry, it is difficult to change a client’s entire desktop to suit your app,” Wooster explains. “Instead, most vendors need to be good desktop citizens – making it easy to be integrated with other components.”

Wooster goes on to add that “mature vendors will have developed many generations of software products over the years. Transforming those to a more modern web-based and micro-application architecture takes time, and you can’t simply turn off the lights while you do the work.”

Continual learning

The patterns and muscle-memory embedded in users’ actions reflect expertise gained over time and can be of critical importance to application design and ergonomics. This user action data can be harvested to develop machine learning-enabled desktop analytics which drive customisability, enhance user productivity, and reduce user error. Implementing machine learning driven analytics is not a point-in-time, one-off activity, however.

To unlock the full potential of machine learning, firms must implement a continual learning regimen which forces models to learn continually – i.e., autonomously learn, adapt, and refine. Continual learning is a key aspect of machine learning model management that formalises how to monitor, manage, and mitigate model drift, thereby ensuring that models are current.

Glue42’s Desktop Integration Platform is seamlessly integrated with the Glue42 Server, an optional back-end component that enables centralization and management of common data and services. The combination of Glue42 platform and server can facilitate the implementation of powerful continual learning capabilities in conjunction with industry-leading machine learning platforms.

The net of it

Glue42 is uniquely positioned to revolutionise the financial service professional’s desktop. Glue42’s robust integration capabilities, combined with Glue42 server can minimise desktop fragmentation, power low-code/no-code driven development, and provide a platform for continual machine learning-driven user analytics, behaviour optimisation, and knowledge sharing.



New GFT Engineering blog on Medium

Other | Artificial intelligence New GFT Engineering blog on Medium

Machine Learning in Credit Risk Modeling

White Paper | Artificial intelligence Machine Learning in Credit Risk Modeling

Model Risk Management in Banking: the big challenges ahead

Other | Artificial intelligence Model Risk Management in Banking: the big challenges ahead

10 Measures of Automation Maturity in Financial Services.

Best Practice | Artificial intelligence 10 Measures of Automation Maturity in Financial Services.

Appian Europe Ltd