Skip to content

Taking Low-Code to the Next Level

A significant amount of both media attention and venture capital money has recently been given to low-code and no-code tools that focus on the areas of analytics, data science, and artificial intelligence. This blog will review the value propositions of low-code tools, discuss why the low-code trend really isn’t new, and will then discuss how some companies are taking the concept to the next level. This topic is also highlighted in IIA’s 2022 Predictions and Priorities webinar and research brief.

Reviewing The Low-Code/No-Code Value Propositions

There are several core value propositions of low-code/no-code toolsets. The one that typically gets the most attention is the way that these tools enable a broader spectrum of people to make use of analytics through the use of point and click environments that they are comfortable with while allowing them to avoid having to actually write complex code.

Equally important, and often undervalued, is the way that low-code/no-code toolsets enable higher levels of efficiency and consistency for experts. When experts can quicky grab pre-packaged, pre-tested objects and easily drop them into their own processes, it not only saves them coding but also ensures that they are applying logic consistently.

Bringing these first two benefits together leads to a third core benefit. Namely, once an organization has a broad range of users from not-so-expert to deep expert using the same tooling, it becomes that much easier to debug, deploy, and support analytics and AI processes. Not only can experts help debug and enhance processes built by non-experts, but the experts can push advanced objects and functions that they create directly out to the non-experts so that they can make use of them. The entire organization becomes more consistent in its approaches, and pushing deployments from a single (or small set) of tools is easier to manage than deploying from many.

Low-Code Really Isn’t A New Concept

For all the hype around low-code/no-code tools today, the tools are just building on a decades-old trend making use of all the technologies available today. In the early days of computing, programming was done in assembly language. I recall trying to code a bit of assembly language years ago and realizing how painful and difficult it was. Languages ranging from Basic to C to Cobol arose in the early days of computing in order to allow much more friendly syntax to be used to request complex functions from a computer. What we now consider “code-heavy” languages were actually only low code relative to the other options available.

Consider the Base SAS programming environment that has been a staple of analytics organizations for many years. SAS has built a lot of visual low-code tools in recent years, but even Base SAS is a form of low-code tool in its own right. With the broad range of SAS Procedures that were created, users could request highly complex analytical and statistical computations to be executed through the use of a well-documented procedure call while setting the desired options. Sure, there was still a lot of coding involved. However, there specifically was no need to code the fundamental (and highly complex) analytical algorithms themselves. It is far easier to call a SAS Proc Reg than to code your own regression algorithm.

Today, with mature graphical user interfaces and a diversity of computational environments, it is possible to build interfaces that allow users with little or no coding knowledge to call those same complicated procedures that analytics professionals used to have to write code to access. Of course, there are some risks with the democratization of such algorithms as I have discussed in the past here. But fundamentally, today’s low-code/no-code trend is simply using all technologies available today to make tools even easier to use. In addition to core analytical algorithms, low-code analytics and artificial intelligence tools today are also incorporating data access, data manipulation, process deployment, and process monitoring functionality. In effect, these tools combine a wider range of functionality while adding an easier-than-ever interface on top of it all. This is all great progress that continues to build on a long-term and well-established trend.

Next Level No Code Functionality

Classic CRM and ERP tools have for years had embedded analytics that include a wide range of analytical functionality and reporting. However, general enterprise tools traditionally included the additional functionality within special menus or modules that usually only a limited, more sophisticated user base could access. Traditional CRM tools largely focused on fairly basic, “slice and dice” segmentation methods where users applied simple filters to standard fields. Enterprise software is now enabling even analytically unsophisticated users to take advantage of cutting-edge analytical and artificial intelligence processes seamlessly and, in some cases, without even realizing it.

A great example of this type of next level functionality is Adobe’s Sensei. Adobe recently demonstrated various ways it is embedding AI for its non-technical users to make use of. Many of Adobe’s core users are specifically non-technical artists who don’t know a thing about AI (and likely don’t care!). Sensei allows these users to easily select objects within a picture, shift backgrounds, or even create a video by interpolating frames between two images taken a short time apart. Users realize the benefit of sophisticated AI by simply pointing and clicking as they usually do. These users don’t need to realize that AI is behind the scenes handling tasks such as identifying the various objects within a picture and their boundaries so that the objects can be selected and manipulated. The users just love what they perceive as cool new features!

Embedded functionality like that in Adobe Sensei takes the low-code/no-code trend to the next level by allowing people to use AI and analytics processes as part of their routine workflow without having to understand how the underlying technical processes work – or even that they exist at all! Many core Adobe users have no interest in building analytics or AI processes and wouldn’t necessarily think that they have a need for them. However, when tools like Sensei take low-code to the next level to provide embedded functionality built on top of analytics and AI, they love what they see.

Originally published by the International Institute for Analytics