A 2023 survey conducted by Microsoft found that 87% of CEOs believe that increased AI and automation embedded into low-code platforms would help them better use the full set of the technology's capabilities.15 Industry experts, meanwhile, are seeing a world where the rise of GenAI makes the need for low-code platforms more critical. For example, leading independent market research company Forrester paints a picture where low-code platforms become the basis for a new platform category: application generation platforms.
“Application generation (AppGen) platforms represent the evolution of practical platform engineering to take full advantage of AI (especially generative AI) while mitigating its drawbacks. AppGen platforms will integrate the steps of software analysis, development, security, testing, and delivery by providing TuringBots for both low-code and high-code development spanning every step—all while incorporating the principles of agile and DevOps along the way.” 16
—John Bratincevic, Principal Analyst, and Diego Lo Giudice, VP, Principal Analyst, Forrester Research,
Low-code platforms are full of AI features, and they are well-equipped to take advantage of new GenAI capabilities by adding governance, definition, and security to the process. Currently, GenAI tools can only produce code suggestions, blocks of code, and small modules. Therefore, developers must still evaluate the code generated to adjust interfaces, understand boundary conditions, and evaluate security risks. Low-code platforms automate that evaluation, so when the code generation is combined with the code review, there is less work for the developer. It is also easier to get started with GenAI when you use a low-code platform because it does so much of the hard work of GenAI development up front so humans don’t have to.
Let’s look at a few ways low-code platforms can help harness the power of GenAI-produced code without the risks, all while providing the support your team needs throughout the entire software development lifecycle.
Get GenAI capabilities plus the infrastructure and tools for the entire software delivery process
Even if GenAI-produced code was perfect, it still wouldn't be enough to deliver a working, validated, and secure application. For one, GenAI-produced code is a black box, so you don’t know where the code is coming from or what’s in it. Also, GenAI-produced code does not know your infrastructure—your systems of record, your databases, or how everything is integrated. GenAI-produced code will not have information about the modifications or extensions relevant to your landscape. It can point you in a direction, yes, but you need specific context for your infrastructure.
“The thing is, writing code to build an app is just a small part of the development process. Everything that comes after is a lot more difficult, and GenAI alone cannot manage the complete software lifecycle and continuous development right now.”
—Elton Escaleira, Product and Service Manager, Bosch
But if you start searching for a GenAI tool that can handle production and post-production, you will soon find that you will need to use a collection of tools—some of them proprietary. Yes, it’s the same old refrain—lots of tools, which means developers have to switch between them.
You can avoid that refrain. With low-code, you can use GenAI in a platform to produce apps and then manage the rest, from building to testing to deploying to maintaining. It also provides the core infrastructure needed to run a modern application portfolio, which costs millions of dollars to stand up. You can take GenAI-produced code and develop applications that are integrated with existing systems and databases. Version tracking, dependency checking, and impact analysis are all part of the package.
“The DevOps capabilities, like one-click deployment and performance monitoring, mean the apps we deliver this way are less costly to update and maintain.”
—Tony O'Halloran, Total Produce
Implement GenAI-produced code with guardrails for security and governance
Artificial intelligence can easily replicate security issues from user codebases and open-source projects, and it can pull in code libraries that are owned by other entities, violating IP. When GenAI is part of a low-code platform, however, strict guardrails and backup checks from central IT rein in the potential for non-compliant applications. A comprehensive approach ensures that any GenAI-produced code with issues is identified quickly and corrected, assuring its inherent security.
“In 2022-2023, the rise of AI Assistants is strongly correlated with 'mistake code' being pushed to the repo… If the current pattern continues into 2024, more than 7% of all code changes will be reverted within two weeks, double the rate of 2021.” 17
Because they have a suite of security and compliance tools built-in, low-code platforms provide automated validations and enable apps to meet global standards like ISO and SOC. Automated security assessments enable them to ensure that applications adhere to stringent quality assurance protocols while AI capabilities constantly review code to check for issues. In addition, fixes for distributed denial of service, newly identified code vulnerabilities, mobile threats, and other protections are automatically applied to your apps.
“Developers are using generative AI alongside tools such as low-code to create applications at unprecedented speeds and do more with the same resources. These technologies’ built-in guardrails foster experimentation while eliminating the privacy and security risks associated with public AI models.”
—Sílvia Rocha, Vice President of Engineering, OutSystems
GenAI also increases the likelihood that non-developers will create a proliferation of applications that are not vetted or duplicate those that already exist. Low-code platforms address that in a couple of ways. For one, they offer tools for managing version control, releases, and component dependencies so non-developers can't create apps that are similar to or conflict with existing ones. The other way is that they provide a single platform so all the developers use the same tools and processes.
“We wanted a secure platform that would allow us to build the capabilities we needed quickly and easily—and that’s the beauty of working in a low-code way.”
—Karel Nouwen, Medtronic
Deliver and maintain GenAI-produced code predictably and consistently
When AI generates code, it's like a black box with no clear owners. We don't really know why it chooses to write the code the way it does. Its source is the vast repository of existing imperfect software written and copied and pasted by humans, who make plenty of mistakes. One example is the coding copilot, a GenAI companion that helps a developer generate code that is, in reality, 50% longer than it would be if written by hand. Generating code that shouldn’t be written increases technical debt.
“In the software development arena, people want to understand why. When they write it, they understand why, but when they generate it, they don't necessarily understand why. The why is very important.”
—Paulo Rosado, Founder and Chairman of the Board, OutSystems
There are other common issues. A copilot can generate an application that doesn’t adhere to defined standards that simplify maintenance, and you cannot easily reuse and modify existing code. All this makes it hard to be sure that the code follows rules and best practices. It can also be tricky to debug and refactor GenAI-produced code when needed.
“Basically, humans find it hard to trust a black box—and understandably so. AI has a spotty record on delivering unbiased decisions or outputs.” 18
—Donncha Carroll, Lotis Blue Consulting
By using a low-code platform with GenAI-produced code, you benefit from speed and ease while still ensuring that your codebase remains maintainable and that you can see it. The low-code platform provides a structured environment to control how AI-generated code is incorporated, making it easier to understand, debug, and evolve that code as needed. This allows you to use the power of generative AI while mitigating some of its inherent risks and challenges.
A low-code platform has built-in features for testing tools, CI/CD, monitoring, and user feedback management. Visual debugging features can pause execution at breakpoints and run logic step-by-step to find errors. Quality analysis, automation, and AI deliver sound apps and architecture. Real-time monitoring, testing, and feedback keep their performance and experience smooth. A low-code platform also makes it easy to execute rollbacks and monitor infrastructure and applications—as well as adjust code as technology and customer preferences change.
“We see orders of magnitude of compression for organizations looking to build new projects and want to release them in 4 months, and it's pretty complex. But, suddenly you can do it with a relatively small number of people and a lot of help from GenAI.”
—Paulo Rosado, Founder and Chairman of the Board, OutSystems
These are just a few examples of why the future of low-code is brighter than ever before—and why combining it with GenAI is a natural progression in its evolution.
But, no matter what method you plan to use, you need a strategy that focuses on the future when you start building GenAI apps.
15 Richard Riley, Low-code signals 2023. Microsoft, 13 Apr.
16 John Bratincevic and Diego Lo Giudice, “The Rise Of Application Generation Platforms, 2024. Forrester blogs, 7 May.
17 Coding on Copilot: 2023 Data Shows Downward Pressure on Code Quality, 2024. GitClear, 26 Jan.
18 George Lawton, AI transparency: What is it and why do we need it?, 2024. TechTarget, 25 Jan.