Low-code and no-code development gets a makeover as priorities shift to AI

Coding abstract in cube form

Yaroslav Kushta/Getty Images

The low-code and no-code market is big and will get even bigger. What’s more, the addition of AI-based assistance to these tools could lead to even greater market growth.

The low-code and no-code market is worth $13.2 billion globally, with a growth rate of roughly 21% annually since 2019, according to research by Forrester analyst John Bratincevic. He says this growth stems from “the institutionalization of low code in IT,” with 87% of enterprise developers working with low-code and no-code tools or platforms. Citizen developers will triple the size of this market by 2030, he continues: “The democratization of development to workers outside of IT shows no signs of slowing down.”

Also: How to use ChatGPT to write code

AI is the key factor that could help to further accelerate this market — up to $50 billion within the next four years. AI will lead to more citizen developer involvement, Bratincevic states. Conversely, he adds: “AI-infused development platforms (TuringBots) could make traditional high coding so productive that professional developers reject low code and switch back to high coding everything.”

AI’s impact on the course of low- and no-code development might fall somewhere in between, he states — with healthy growth fueled by the integration of AI and low- and no-code platforms.

Also: Okay, so ChatGPT just debugged my code. For real

Still, it should be noted the capability to deliver AI applications with low-code platforms — assuming the citizen or professional developer is ready to do it — can be problematic. High levels of development skills are still needed. “The language that is used to develop gen AI is not really English,” Rodrigo Coutinho, co-founder and AI lead at OutSystems, pointed out in a recent podcast hosted by Amazon Web Services. 

“You do need to learn the dialect,” he urges. “Right now, you need courses in engineering, because you need to learn in a way that the machine can understand what you’re saying. Even though the words are the same, you don’t really talk to it as you would to a person.”

This language barrier may impede the long-heralded democratization of software development, he continues: “It’s not as complicated as C# or JavaScript. But it’s a language that you need to learn in order to be able to develop.”

It’s also important to note that there’s still a major distinction between AI-assisted development and low- and no-code development. “Gen AI brought a huge boost in productivity for traditional developers, but they still need to know what they’re doing,” says Coutinho. 

“To use gen AI tools for traditional code, you still need to be an expert. Even though a lot of the work is done by the machine, you still need to be able to read the work that was created, understand it, adapt it to your own needs, and change it. The first version of the application is just that.”

Also: How to use ChatGPT to create an app

Therefore, it’s probably too soon for inexperienced developers to work directly with generative AI to build applications, agrees David Isbitski, principal developer advocate at Amazon Web Services, who joined Coutinho in the podcast. Unless you are familiar and experienced with programming, “you don’t know what you don’t know.” AI-enabled development takes not only technical experience, but also a sense of what and how the code needs to be mapped to the business process. 

“If you’ve been coding for some time, you know as a human being how to do a process,” says Isbitski. “You can change that process into code. But someone who hasn’t written software before wouldn’t know what to ask.”  

As AI does enter the development workstream, the technology might be tapped as an empathetic assistant. The ideal AI assistant within a low-code environment “can analyze my thought process,” says Isbitski. “This is how I wrote this code, this is what it will do, and this is how it’s been enabled. It seems like magic. It’s that encouraging and making sure things are correct.”

Ultimately, an ideal AI assistant can better understand the context in which software is being written and deployed, he continues: “As human beings, we know all these things, what day it is, the climate of the world, that AI hasn’t known. These are things that are important to the output. Bringing that stuff back, as you’re having these conversations as you’re writing code, is incredibly powerful.  

Also: How to write better ChatGPT prompts

The goal of AI-assisted software development is to “allow people to learn and improve themselves,” says Isbitski. “Instead of just giving people the answers, give them the chance to come to answers themselves. It’s incredibly powerful as a teaching tool. Maybe because these gen AIs and LLMs are truly a reflection of us.”

Ultimately, AI-infused development should lead to fresh opportunities for developers. “In leading teams, a big chunk of their jobs is to look at code by their juniors and make sure it’s okay, meets the requirements, has quality, and so on,” says Coutinho. “That’s one of the impacts that gen AI will have on the life of the developer. The individual contributor is the gen AI, and you are the team lead that will make sure everything is alright.”  

Leave a Reply

Your email address will not be published. Required fields are marked *