Don’t Let Your Company Get Left Behind in the Drive to Low-Code/No-Code Adoption
What’s driving the adoption of low-code/no-code solutions, and what do you need to know to implement it in your own enterprise?
- By Mike Loukides
- July 17, 2023
With the number of tech layoffs increasing and generative AI advancing rapidly, companies are under intense pressure to adapt by integrating these tools into their daily workflows. Meeting the demands for increased productivity amidst a shortage of IT skills has created a need to improve the skills of both developers and non-developers in AI-based systems.
The emergence of low-code and no-code tools has already provided companies with a solution. These tools enable employees of all skill levels to become more productive in creating software, internal reports, and legal documents, making them an invaluable asset in navigating the IT skills shortage and driving success in the face of economic challenges.
There is also much room for growth as more companies make greater use of these tools to streamline tasks and create automated workflows to increase productivity and efficiency. However, there are risks in forging ahead blindly with low-code and no-code tools. Companies need to be sure that the citizen developers enabled by the tools are working with IT teams to ensure that the apps they create are efficient, properly deployed, and secure.
In a recent report from O’Reilly Media, respondents indicated a growing interest in artificial intelligence–assisted low-code and no-code tools, revealing how the tools are being used and why. The report also pointed to a larger need for more training around AI-based systems, particularly as tech layoffs increase and employees see the need to grow their arsenal of competitive tech skills. Key findings of this survey reveal that data leaders and their teams can gain a deeper understanding of the implications of low-code/no-code productivity tools in the workplace and their long-term effects on developers and non-developers alike.
Drivers of Low-Code/No-Code Adoption
Increased productivity. The common thread for why low-code/no-code tools are used boils down to productivity. Organizations limited by their current staff still need to produce more solutions in less time. When asked why they needed the tools, 15% cited long times between ideas and outcomes, 13% said they needed more creative solutions, and another 13% cited a growing need for unique solutions that address specific needs. Twenty-two percent of respondents said low-code and no-code tools saved time by writing boilerplate or scaffolding, and another 10% said these tools enabled them to develop and test more alternatives.
A wide range of uses. The survey found that 44% of respondents are currently using these tools for applications ranging from code generation (20%) and business analytics (13%) to generating marketing content (4%) and design (4%). Although less than half of the sample use the tools now, 35% said they would like to. Only 21% said they don’t plan to.
Closing the skills gaps. At many companies, multiple rounds of layoffs have crimped their ability to hire. Although hiring didn’t stand out as a major issue for respondents, 10% said these tools helped them fill a skills gap without hiring. Furthermore, these tools also help improve the skills of those on staff. Nearly one-quarter of respondents (23%) said low-code and no-code tools helped bring beginners and entry-level staff up to speed quickly. They also reported that the tools can help employees go from mid-level to expert (11%) and from expert to superpower range (7%).
Humans Own the Results
Companies may not be focused on using low-code and no-code tools directly to mitigate talent shortages, but the tools unquestionably save time by allowing developers and non-developers to drag and drop components into applications. Allowing business analysts, managers, or others to create applications without knowing how to write code can significantly shorten development times.
However, there is more to using these tools than simply dragging and dropping. Organizations need to think about what they are using low-code and no-code for and who is using the tools. They also need to consider the requirements for deploying efficient, reliable, and secure applications. AI projects necessitate significant work to structure an organization's data and prepare it for AI training.
At one end of the spectrum, applications such as ChatGPT can be used to help developers write code. They're proving useful -- if not always flawless -- but they're better at giving senior developers superpowers than they are at helping junior developers. Although the tools are good at programming language syntax, users still need to read the code to understand whether it’s correct and debug it when it isn’t. Users also need a thorough understanding of the problem they’re trying to solve, and they need to write detailed prompts describing the desired solution. Superficially, it looks like programming just became easier, but software developers need all the skills they already had, plus new skills to use these tools effectively.
There also are plenty of low-code/no-code tools that can be used for building websites, business intelligence dashboards, and spreadsheet models for working with databases. These tools open the door to non-developers because they often don't require the user to write code at all. As with the code-writing tools, however, the user needs a thorough understanding of the desired end product.
If you’re building a website, you need to have a good feel for how the site will look, how it will function, how to make it easy to use, and so on. A tool can handle programming tasks involving HTML, JavaScript, and cascading style sheets (CSS), but you must still make the hard design and functionality decisions.
Merging Low-Code/No-Code with Security and IT
It is difficult to know the security risks these tools introduce because they involve code generated within a third-party platform. Citizen developers don’t have visibility into the quality of code, how it conforms to security best practices, or how much security monitoring was done during development. Furthermore, organizations should be concerned about performance issues, testing, and deployment models.
These are all classic IT issues that won’t go away with the use of ready-made code. A manager may be able to build a model using a tool for automated machine learning, but she can’t assume that it will be free of bias, that it won’t get stale over time, and that it will behave as expected. It is essential to remember that humans are responsible for the tools they create, even when those tools are “created” by AI. Tools can’t be held accountable for their output. Humans always own the result.
Broadening the number of people who can build usable software without having programming skills will clearly save time and effort for companies trying to make the best use of their employees. We're still trying to define the proper relationships between these developers and more formal IT processes. Will a sales manager build a better web app for managing the sales team than a group of IT developers who rarely talk to the salespeople? Probably. Can that same manager deploy the app, verify that it's secure, and do all the other things we expect of IT teams? Probably not. Before going too far down the low-code/no-code road, we need to figure out how to make those relationships work.
Ironing out the details of how these tools will be implemented is important because they are here to stay. Within a few years (if not months), low-code/no-code won’t be an exotic new technology; it will be a standard fixture in the workplace. Companies that want to create better workflows and streamlined processes will need to prioritize low-code/no-code adoption. In due time, if you’re not using those tools, the only question will be why you’ve chosen to be left behind.