AI is ubiquitous. If not in active use, then in conversations with colleagues at work or at the family celebration when everybody sits at the table. But well-founded knowledge of what is being talked about is rarely present. So, it’s not that surprising that many companies actually would like to implement intelligent technologies in their businesses and processes, but lack the personnel or access to the skills that would be necessary to do so. In this short article, you will find out why this is not really a big deal and why it is still possible to implement such technologies.
Who’s not familiar with the following: Actually, the will would be there to do something where you know that a significant benefit would be the result. But then you find a suitable excuse that sounds quite plausible to you and you use it to justify that you let it be after all. Does that ring a bell at all? If so, be reassured. You are not the only one…
At least in the context of the integration of AI into business processes, this seems to also apply to a large number of German companies. According to a survey by IDC, it is especially the lack of skilled respectively knowledgeable workers in the company that prevents management from launching such projects. In itself, one might think, this is a somewhat plausible reason for doing nothing. However, the following points illustrate why these companies are wrong:
Thanks to AI technologies from the cloud, there’s no need for any in-house infrastructure
Cloud services have become increasingly popular in recent years and have established themselves in a wide variety of markets. Today, there is also a large selection of different providers who offer you as a customer an out-of-the-box solution, where you get not only the software and interfaces, but also various services on top of that. In short: You get the necessary infrastructure without any in-house experts to set it all up. You can therefore scale the applications for your use cases individually and according to your individual needs without having to take the necessary precautions. Anyhow, it is important that you inform yourself beforehand about the regulations of the operating options and clarify whether the existing APIs are compatible with the IT systems currently in use at your company. If this is not the case, you should negotiate bilaterally with the provider about the possibility of building such an API.
Of course there are also various cases where a cloud solution is simply not an option. If this is the case, you should first consider working on other potential use case.
I can’t find any data science people – that’s from yesterday
One of the consequences of all the sensation and hype around artificial intelligence and machine learning is that more and more people are educating themselves in exactly these particular areas. So what a few years ago was a big challenge, namely the search for such experts, is today only a question of the offer of employers. However, it should not be forgotten that this is precisely what is difficult for various companies. Since the demand for top specialists is still much greater than the supply, there is enormous competition for talent. For many, the cooperation with external partners is an attractive and reasonable alternative. These partners are involved only on a project basis and support the preparation of company data, the development of algorithms and the launch. As soon as the applications are up and running, you can work autonomously and are only dependent on support again when there are escalation situations.
Practical use cases come from front line of the business
Contrary to the potential assumption, the inputs and conceptions for promising use cases for artificial intelligence applications often come from the business side and not from the IT department or from AI specialists. The reason for this is simple: The know-how for the daily business at the front is in the business units. In other words, the key knowledge can be found among the people who are confronted daily with the different processes of the business activities and have thus built up a context understanding for all possible stakeholder interactions. That’s why a promising approach for the development of application scenarios is when business experts take different processes apart, for example in a workshop framework, and collaborate with the inhouse IT specialists to check for the potential for optimization. In addition to a target-oriented project implementation, you can possibly also break the potentially existing cultural barriers and promote employee cooperation across departments. Who knows what opportunities this will reveal for you and your employees on top of everything else.
As you can see, the possibly missing technical know-how for the implementation of AI projects is definitely a problem you can overcome. What can truly be a problem though, is the lack of qualitatively sufficient data, which could be used to “feed” an AI application in order to generate meaningful and economically viable outputs. However, if you can preclude the lack of quality of your data, then you should now follow your intuition and kick off your AI project. Good luck and enjoy the process!
