Why Low-Code Isn’t the Future of Software Development

Karel Callens, the CEO and Founder of Luzmo, shares his take on why low-code might not be the future of software development. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

Low-code options for building and designing apps are becoming increasingly popular thanks to their ability to sidestep the need to consume developer resources and allow businesses to quickly build tools that are customizable for most use cases. However, as businesses become more complex with increasing specificity now a defining trait of the modern marketplace, the challenge is building templates that are widely applicable enough for a broad range of purposes.

For low-code development to scale, the initial range of options—whether those are templates, charts, graphics, etc.—needs to be adequate for a large customer base. Equally, when selecting a low-code solution, it is difficult for a business to accurately assess how much customization will be needed early in the buying process, particularly as data monitoring needs can be emergent and time-sensitive.

Additionally, as businesses continue to demand technology that can deliver increasingly hyper-personalized experiences, they also need access to the building blocks for systems that offer a higher level of customization than current low-code alternatives. These are tools that can then be used either in-house to enable the quick turnaround of turnkey solutions or sold on to businesses that have the resources and expertise to develop on top of a software development kit (SDK).

To manage this uncertainty and effectively respond to changing demands, a ‘mid-code’ option is emerging in the SaaS marketplace. Mid-code offers a similar level of customization to pro-code tools but handles the more technical aspects in the background to keep the workload realistic for a small or even one-person team that will be in charge of implementation.

What on earth is mid-code?

‘Almost no-code’ doesn’t roll off the tongue quite as well as low or no-code, but it is the most accurate way of describing the new family of data experiences and the tools that build them. Low-code has a huge range of uses, but it is following a transition that is common to most technology and becoming widespread in consumer applications. Instead, software for enterprises is becoming more complex. Consequently, to manage this additional complexity, most applications require developers to add ‘finishing touches’.

These ‘almost no-code’ or mid-code solutions (as I have just termed them!) have broadly the same philosophy as low-code, simultaneously handling the most complex and unwieldy parts of software development in the background while also ensuring that out-of-the-box constants required for most customers (e.g., multi-language support) are provided as standard. Crucially, however, they offer the functionality that moves the end product towards a greater level of customization.

Preparing for future technology and the modern market

As mid-code solutions increasingly meet developers and businesses at the halfway point through enhanced customization, so too are the winds of change ringing across the modern SaaS landscape. Firstly the market for SaaS products is becoming increasingly fractured, which is requiring a higher standard of interoperability to build cohesive end products. Secondly, the apps themselves are rapidly changing. The reason? Yes, you guessed it: AI.

AI developers have prioritized API integrations as a prerequisite for commercial success. As a result many platforms are already using these APIs to directly embed AI functions into products.

APIs have always been integral to the SaaS ecosystem, allowing the patchwork of different vendors to interoperate and collaborate. Putting APIs at the center of development enables the degree of modularity needed for the level of customization most customers require.

This level of interoperation and tech agnosticism is a recent development. Throughout the industry, certain programming languages and API protocols have coalesced into a series of standards that can realistically support interoperation between different data sources, interfaces, and query languages. Now that these tools are available mid-code is a more realistic proposition for developers and customers.

Pro-code is changing too 

With the way mid- and low-code applications are changing the market, pro-code and the practice of coding are undergoing a seismic shift. Some industry leaders are suggesting that coding itself will change as a profession, even warning new graduates against training in a traditional developer role. In part, this is because predicted developments in AI will radically change the process of writing applications.

Despite this gloomy prediction, data from GitHub suggests that there are more developers than ever, with the site’s user numbers having more than doubled in the last five years. Similarly, Microsoft CEO Nadella suggests that every business is a software business on some level. So, where is the actual role of developers going to end up?

It’s likely, if not already the case, that developer skill sets will fracture along the same lines as software development. Some developers are able to build applications from the ground up, some with specific expertise in data management or distributed systems that can lean on other tools to make up for any skill deficits, and then, finally, low-code developers. The term seems like an oxymoron, but the ability to quickly build effective apps with low-code tools is a kind of know-how that will still add value to the after-sales experience for customers.