The Rise of AI-Infused Applications: Transforming Software Development

Today, I'm joined by Koen Janssens (Global Technology Lead for Applications) and Kim Rymenants (Global Sales Director) to discuss AI-infused applications and the transformative shift in application development that lies ahead. We talked about:

• Why chatbots are just the tip of the iceberg ...
• ... with AI-infused applications being the iceberg
• How LLMs can change how we design software ...  
• ... that allows non-tech users do very heavy tech lifting
• And what all of this means for data-driven businesses!
  

  
  Let’s roll 

Kim, Koen: Welcome! We all know 2023 was a big year for AI, especially with advancements in Large Language Models (LLMs) and their use in chatbots. 

Kim: It was a time of rapid innovation and exploration. You're right, chatbots became a major focus, making user interaction more engaging and accessible. At Cegeka, we embraced this by developing our own chatbot framework, Milo, to speed up custom chatbot creation. But  chatbots are just the tip of the iceberg, to be honest.  

Which begs the question: what's the iceberg then? 

Koen: The bigger picture is what is called "AI-infused applications." These are essentially apps with built-in AI capabilities. Industry experts like Gartner predict that by 2026, around a third of new applications will leverage AI, compared to less than 5% today. This shows a shift towards software that's more intelligent, responsive, and user-centric. 

"Industry experts predict that by 2026, around 30% of new applications will leverage AI."

Kim: And it's already happening. All our customers rely heavily on data, and for some, data is their core business. They're discovering LLMs can unlock more potential from that data than ever before. For example, LLM technology can be used to fully automate and AI-enhance complex workflows, saving significant time and effort.  

Koen: Imagine an LLM-powered application that guides non-technical users through an intricate technical process using an easy-to-understand interactive wizard. This wizard analyzes data and automates tasks in the background. Result: non-technical users can tackle challenging technical tasks without needing expert help, reducing weeks of work to just a few hours. 

That’s quite a tall promise. Can you give a specific example? 

Koen: Sure. Here’s one that is applicable in every sector of industry: uploading large amounts of raw data into a system and ensuring it's converted properly so the system can immediately work with it. It can be HR uploading huge data sets to their HR service provider portal, or marketers uploading data to CRM or marketing tools, or any other similar scenario. 

With AI, this process can be entirely automated. Previously, it required a data engineer's manual effort and time to ensure the data was formatted correctly. Plus: users spent weeks sweating over data files. Now, AI-infused wizards can help non-experts clean, convert, map, and upload large datasets in just a few hours, easy step-by-step. A bonus is the improved data quality by eliminating human error.  

Kim: And data engineers get to do more exciting stuff (laughs). Koen’s example may seem  simple or straightforward, but it addresses a real and recurring pain.  

"AI-infused wizards can help non-technical users clean, convert, map, and upload large datasets in just a few hours."

How will AI impact software design? 

Koen: The expanding range of AI tools means we need to approach software design differently. We have a much bigger toolkit now. As designers and architects, we need to think beyond traditional methods and leverage these new tools. LLM technology isn't limited to just chatbots; it can be applied creatively in various ways.  

Kim: It's a whole new world with countless business cases and endless possibilities, some of which we're only just beginning to grasp. If you break down an application into its building blocks, there's likely an opportunity to use AI to speed up, simplify, or automate tasks within each block. That’s how we need to start thinking.  

So, this is especially relevant for data-driven businesses?

Kim: Absolutely. Businesses where information is the lifeblood of their operations stand to gain the most from this approach. They need high-quality, user-friendly interfaces for their customers. Their true competitive advantage lies in augmenting services with AI. This can apply to customer portals, core business applications – essentially, the central nervous system of their operations. 

"If you break down an application into its building blocks, there's likely an opportunity to use AI to speed up, simplify, or automate within each block."

Does this require a complete overhaul? Can we add AI capabilities to existing applications? 

Koen: Yes we can. We can think of existing applications as a collection of building blocks. These blocks can be swapped out or enhanced with AI-infused modules. It's not always a simple process, but absolutely achievable.  

To conclude: you mentioned chatbots, and AI-wizards. Any other exciting AI projects you want to highlight? 

Kim: At Cegeka, we're particularly excited about projects that drive measurable sustainability benefits. A great example is our work with digital meters. These devices typically send data every 15 minutes, but occasional connectivity drops create data gaps. By leveraging AI to analyze factors like temperature, holidays, and historical usage patterns, we can fill these data gaps. Consumers receive accurate, real-time energy consumption estimates, even when the meters fail to communicate. This empowers us all to make informed energy choices.