Artificial Intelligence within the IT: the benefits and its potential


In the last couple of years, Artificial Intelligence (AI) has drastically and permanently impacted our way of living. Think about face recognition that we use to log into our devices, the recommendation systems in social media like YouTube and Tik Tok, or the small autonomous vacuum cleaners and lawnmowers in our homes. The recent release of OpenAI’s ChatGPT has shown us that AI is here to stay and has caused tech companies to integrate AI-systems into their IT-solutions. With the availability of numerous vast data volumes and the use of powerful computer hardware, incorporating AI into your IT-landscape has never been this easy. However, what is exactly its place within the IT-world?

Growing Influence of AI

Before diving into AI and its impact, it is important to define a few concepts and to distinguish the difference between AI and Machine Learning (ML). The term ‘AI’ is an umbrella term for all computer systems (often called ‘AI-systems’) that are capable of performing complex tasks requiring human intelligence, such as decision making and prediction. ‘ML’ forms the biggest subset of these AI-models, aiming to find relationships and patterns within a dataset through statistical means. Each Machine Learning model is built (or ‘trained’) for a specific task, such as predicting an amount of supplies (demand forecasting) or classifying customers into subgroups. Since Machine Learning forms the biggest subset within AI, both terms are often used interchangeably in media. This is valid since Machine Learning realizes what AI conceptually does: emulating human intelligence. Therefore, in this blog, whenever the term ‘AI’ or ‘ML’ is coined, we mean the Machine Learning model implemented within the AI-system.

Nowadays, the presence of AI-systems has never been more prevalent in society, being solidified by the release of OpenAI’s ChatGPT service. Companies are successfully adopting ML-models in their business cases, such as recommending personalized products (Netflix, YouTube), or detecting fraudulent transactions (PayPal). However, the recent hype surrounding AI-systems has not been of all times, as companies were skeptical of the benefits AI could bring at the time. As an example, SAP had stated in the past that AI would not be broadly incorporated into their solutions, implementing only some small applications within their Business Technology Platform (BTP). Nowadays, driven by the increasing adoption of AI, SAP has considered AI to be a core part of their business strategy, embedding AI functionalities throughout their IT-solutions, including the BTP (SAP AI Core and SAP AI Launchpad), their business solutions (collectively called Business AI), and data and analytics solutions (PAL and APL libraries in the HANA database).

AI Blog part1 p2

Now that we have a clear view of AI and its adoption in society, how does it fit within an organization and what are its benefits?

Benefits and potential

One advantage of incorporating AI into your IT-solution is that repetitive and time-consuming steps in your business processes can be automated, leading to cost reductions and saving up resources. This allows companies to shift their attention to more valuable work, increasing the company’s efficiency and productivity. As an example, SAP Build Code is a generative AI solution that make use of Joule, SAP’s digital AI-assistant, to generate code in Java and JavaScript. This streamlines the lifecycle of business applications, as you can tell Joule in natural language to generate the necessary code, instead of manually coming up with the correct code. In other words, a developer of the business application only needs to validate the generated code on its correctness, saving time on the implementation of the business application. As such, one can focus more conceptually on the business application and its necessary functionalities.

Another benefit of AI-systems integrated within an organization’s IT-landscape is that it allows the business to react to rapid changes of the market at real-time. This is because numerous and vast volumes of data from various industries (e.g. retail, medicine) are publicly available on the internet, providing a wide variety of information for any business case. Due to their mathematical definitions, AI-systems are able to process these large data volumes efficiently and effectively. On top of that, the availability of high performance computer hardware, most notably the Graphical Processing Unit (GPU), allows Machine Learning models to perform incredibly quickly. This in turn streamlines the development and maintenance of Machine Learning models (i.e. Machine Learning models can be trained in a few minutes). As an illustration, suppose a grocery store is using Unified Demand Forecast (UDF), a module within SAP’s Customer Activity Repository (CAR) solution used for demand forecasting, to predict the amount of required supply for, for example, cucumbers for the next week based on some historical data. After a few months, the dataset contains new information as time passes, such as new sales information or changes in the weather. The UDF module can then be used to rebuild the underlying Machine Learning model based on the updated dataset, taking the new changes into consideration when predicting supplies for the future (see image below for a schematic overview of this process). By regularly (e.g. weekly) performing this update cycle, the grocery store can react in real-time to customer preferences.

AI Blog part 1 p3

Closing Remarks

Today, it has never been easier to make your organization more agile with AI-systems in this fast-paced market. The potential of AI-system is not limited to the most complex models; even the simplest Machine Learning models (e.g. Linear Regression, Logistic Regression) can enhance the efficiency and productivity of your business processes. However, careful consideration is still needed in order to correctly release the solution, as there exist some pitfalls when implementing AI-systems. Stay tuned for the next part where we will dive deeper into those pitfalls.

We're here to listen.
Get in touch with us.

About the author

Photo of Chris Al Gerges
Chris Al Gerges

Chris Al Gerges is a BI consultant at Expertum.

Read more articles by Chris Al Gerges