Low code: AI is lowering the barriers to entry, report

Low code tools are starting to evolve beyond their traditional sweet spot

Low code development tools have been around for quite a while, but they never seem to have quite broken through in the way they promised

Perhaps the label "low code" (or even "no code") oversells them a bit, suggesting that they can replace skilled developers in the creation of complex business applications. In reality, their sweet spot is with basic data-driven web and mobile apps for managing inventory, self-service portals, prototyping, answering customer queries, basic process automation and the like.

But in the right context they can provide faster development cycles, reduced need for coding expertise, and quicker iterations than traditional coding. And of course the tools themselves continue to evolve, with additional integration capabilities, options to incorporate custom code, and AI assistance in the shape of copilots.

Computing has covered a few low code success stories over the years, including organisations that have built fairly complex applications using a platform approach, but we don't see as many of those case studies as we might.

Will we see more in the future? Perhaps so, according to a survey of 250 IT leaders at medium to large UK organisations, commissioned by service provider Evoke Technologies.

Three quarters of respondents expect generative AI to "turbocharge" low code development, with 73% saying that gen AI will lower barriers to the use of low code development tools.

Evoke is a service provider with skin in the game, so such results should be viewed with caution. But general purpose chat tools are already capable of generating large chunks of code and summarising indecipherable documentation, so this is definitely the way things are going.

A common concern about low code tools is security. There tends to be more of a focus on rapid development than on secure coding practice, and tools may lack controls for identity and access management, encryption and permissions. It can be easy for developers to give applications elevated privileges or to ignore compliance requirements, and because the underlying code is abstracted it can be hard to integrate security after the fact.

Indeed, security was among the issues cited by the 48% of respondents to the Evoke survey who said they struggled to get started with low code after investing in a platform. Other sticking points included scalability, vendor lock-in and licence pricing. Hiring experienced developers was the most common strategy for overcoming these issues.

That said, 83% agreed that low code was living up to its promises, 79% said it improves collaboration, and 60% were already using low code platforms and tools to develop mission-critical applications. More than 90% of respondents said low code tools and platforms have decreased application development timescales, with almost half stating that timescales are at least 50% faster.

Twenty-four percent said they were adopting low code tools to enable "fusion development", combining professionally written software with that of community coders.

It would seem, then, that low code tools are starting to evolve beyond their traditional sweet spot. But writing functioning code is only part of the development process. Enterprise applications require scoping, defining, planning, governance, integration and security, inputs that only experienced developers, architects and security specialists can bring.

Computing's DevOps Excellence Awards 2024 showcase outstanding achievement from organisations, personalities and solutions operating within the DevOps space. Enter today!