By Balakrishna DR, Senior Vice President, Service Offering Head - ECS, AI and Automation at Infosys

Traditional software development requires writing reams of complex code to build the desired capabilities in an application. This requires programmers with in-depth knowledge of computer languages, development environments, deployment processes, and testing frameworks. On the contrary, low-code platforms employ visual interfaces and graphical elements that help to define data, logic, flows, forms, and other application artifacts, without the need to write code; these platforms offer reusable components with ‘drag-and-drop' features that a user can link together to create the desired application.

From faster software development and innovation to more productive teams to the emergence of citizen developers, the list of low-code benefits is exhaustive. Naturally, interest in low-code application development is high: in a recent survey of more than 2,000 IT professionals from six countries, 75 percent of the respondents called it a "trend they couldn't afford to miss."  A leading industry analyst predicts the market for low-code development technologies to grow 22.6 percent this year to touch US$ 13.8 billion.

Numbers apart, what low-code really brings to software development is a new paradigm of democratized programming where even non-developers write applications. However, while citizen development sounds fine in theory, it's not that easily put into practice. Here, Artificial Intelligence (AI) based tools could equip non-technical workers with certain skills to help them get started. We see potential for low-code platforms to leverage AI to support citizen developers in the following areas:

Experience Design: These low-code platforms enable the prototyping of user experiences.  AI supports that by endowing the platforms with the capability to hyper-personalize experiences, automatically translate designs into production code by leveraging Machine Learning-based code generators, offer omnichannel experiences across voice, mobile, web, AR, VR, and meta, and extract information from images and videos with the help of vision analytics.

Digital Experiences and Applications: Low-code platforms in this category help in accelerating enterprise application development. When AI is added to these platforms, it provides out of the box intelligence and prediction capabilities, along with AI-assisted development, including suggestions on next best actions. Other features include AI-powered experiences with context-specific data drill downs, understanding natural language used in interactions, AI-based issue predictions, decision assists, predictive intelligence in incident detection, action recommendation, and cluster analysis.

Digital Process Automation and Operations: Built for users working in the areas of business process engines, robotic process automation, and workflows, these platforms can benefit from AI in several ways. For one, they can leverage it to decipher patterns from previous experiences and decisions to make more intelligent decisions in the future. Next, they can use AI in process discovery, mining data to understand points of friction and points of optimization. Other AI-endowed capabilities include prescriptive and predictive analytics, and voice mining for real-time transcription, sentiment analysis, and deciding next best action.

Enterprise Productivity: The purpose of such low-code platforms is to develop applications that improve employee and enterprise productivity. It is possible to enhance these platforms with various AI-based tools, such as conversational AI (conversational bots), AI-enabled productivity apps with prebuilt models for data analytics and display, AI-powered reporting and analytics, human-bot interaction-based workflow generation, and image analytics for recognizing documents like invoices, purchase orders, certificates etc.

Data Science and AI: To this type of low-code platform, built for developing faster AI models, AI can add value by enabling business users/ citizen data scientists to extract insights using the available simple, standard, and conventional models without IT involvement. The other advantages of AI are EDA (auto exploratory data analysis) with augmented insights, outliers, correlations, and bias detection; domain-agnostic and domain-specific pre-trained models related to vision, speech, translation, etc., as cloud services (e.g. AWS Medical comprehend); and gamification of learning and training AI services to make it easier for citizen developers to train and integrate AI services into their apps. Last but not least, AI brings machine learning (ML) with dynamic and unsupervised active learning, best fit model suggestion through automated feature engineering, assisted active learning, and transfer of learning on pre-trained models.

ML models like GPT-3 and CODEX for code generation, and scaled-down GPT-3 cousin, GPT-J, can be used to train domain-specific models using generally available compute resources.

While low-code platforms, with AI, present exciting possibilities, enterprises need to tread carefully as these platforms have limited flexibility. A playbook on how to use low-code platforms correctly is required to avoid unmanageable apps from mushrooming. Enterprises should take a governance-first and build-next approach; additional security and privacy controls should be configured and implemented to prevent data loss, while ensuring regulatory compliance and controlled accessibility and visibility of their data and environments. Once this foundation is in place, AI-powered solutions can take low-code platforms forward into the future.

##

ABOUT THE AUTHOR

Bali is a senior vice president and heads the delivery for ECS business unit of Infosys catering to energy, utilities, telecommunication, media, entertainment and services industries. He also heads the AI and Automation unit for Infosys and is responsible for driving both internal automation for Infosys and providing independent automation services leveraging market leading products for clients. Bali has been with Infosys for more than 25 years and has held sales, program management and delivery roles across different geographies and industry verticals.

Bali spearheaded creating vertical practices, industry consulting group and solutions to deliver differentiated value-added services to clients. In his previous roles, he headed ADM, SAP and Testing service lines for ECS. He was also head of the Bangalore development center and set up our first Global Development Center in Canada. He has managed several large programs for Infosys for various Fortune 500 clients. Bali participates and speaks at multiple industry forums.