Overcoming the challenges of artificial intelligence

18.10.2019 - Technologiepotenziale


However, tapping into the potential of AI is by no means trivial and requires effort and willingness to explore by the companies, the departments and their emplyoees . 

It is well-known that artificial intelligence offers a huge potential for succes. Every company should think about using AI efficiently as a driver for growth and exploiting its evolutionary and disruptive potential to the best possible way. However, there are different approaches for implementing AI in companies. From our experience, one of the most promising approach for a succesful implementation is to “start small and explore” and consecutively “scale and benefit”. 

For this purpose, lighthouse projects are more than suitable. They help to explore the AI potential for the company and make AI tangible. At the same time first barries and challenges become known, problems can be more clearly pinpointed and different solutions can be probed in a “testbed”-setting (“start small and explore”). However, when transforming lighthouse projects to the operative environment associated to the core business, some important things have to be considered. From our point of view, two dimensions are essential for the success of a lighthouse project becoming an essential building block of the enterprise: effort and scalability 

Implementing AI on the one hand requires effort in different directions along a diverse number of company departments, skills and functions, like infrastructure, data management or services. For example, high effort would be associated to cost-intensive operations of the algorithms for AI services, that do not necessarily create direct business impact on the revenue side or the implementation of AI services from scratch with large development efforts. 

On the other hand, scalability is an important performance indicator for AI in terms of extensibility, modularity and adaptability of the algorithms, by e.g. allowing to scale well the amount of required training data with the number of recognizable objects in a computer vision-setup.  

Dependent on these two dimensions, a lighthouse project can be a success or failure. In general, three different scenarios can be observed, starting from a lighthouse AI project (cf. Fig. 1). 

Fig. 1: Alignment of AI projects with three options, starting from AI lighthouse pilots

The first and at the
same time worst scenario is the “guarantee for frustration”. This will happen when the effort becomes too high and the scalability remains too small. In this case, resources are wasted and a derimental output is produced. Also, there is risk that the organisation does not acknowledge the potential of AI due to arosen problems and misses future chances. 

The second scenario are the “questionmark projects”. In these project both, the effort and scalability, are high but the outcome is not exactly predictable. Success and failure are close together, while each of them can come true. For this type of AI project, an individual evaluation needs to take place and benefits vs. costs need to be balanced on a case-by-case base. 

The last and best scenario is to achieve a “success story”. With the charactersitics of low effort and high scalability these projects are the goal for every company. They allow to realize the strengths of lighthouse projects and could lead to transformative initatives of doing business through AI. 

From our experience, success stories (and partially also questionmark projects) share some important guidelines, that have to be considered: 

1. From our experience, the first step is “fail fast”, which means to be willing to experiment in order to achieve the fastest possible validation or falsification of AI ideas based on data.  

2. The second step is to convey the vision and strategy for AI internally. Specifically the AI vision and strategy of your company have to be clearly defined and communicated to all relevant stakerholders, so that everybody understands the potential of AI, the roadmap ahead and is able to act accordingly 

3. The third step ist to empower your organization. In order to implement AI it is important to adapt the organization, processes, team skills and the infrastructure to be “AI ready”.  

Learn more about the implementation artificial intelligence projects in our summary! Download it now and start to benefit from the potential of artificial intelligence!

Ansprechpartner: Dr. Daniel Nowakowski, Project Manager
Mail. d.nowakowski@muecke-roth.de

Zur Übersicht

Vielleicht auch interessant

Fortsetzung der Themenreihe: Loslegen mit Künstlicher Intelligenz (KI) – Anwendungsgebiete im Marketing und Vertrieb

ONLINE SHOP OPTIMIERUNG AM ENDE? WIE KI BASIERTE METHODEN DIE VERTRIEBSLEISTUNG RADIKAL VERBESSERN KÖNNEN „Jeder kann KI erfolgreich einsetzen!“ Viele Online Shop Betreiber nutzen A/B Testings, um ihre Webseiten Performance zu verbessern. Damit konnten sie in den letzten Jahren den Erfolg kontinuierlich ausbauen. Mittlerweile lässt…

Zum Artikel
Kuchentratsch: Neues soziales Start-up im MRC Portfolio

Omas Kuchen schmeckt am besten Kuchentratsch versendet deutschlandweit die Lieblingskuchen aus der Kindheit, gebacken nach dem Originalrezept von Oma. Das junge Unternehmen beschäftigt die wohl erfahrensten und leidenschaftlichsten KuchenbäckerInnen – unsere Omas und Opas – und bietet ihnen damit eine Möglichkeit gemeinsam in einer angenehmen…

Zum Artikel
Schnelles Wachstum bei Start-up Agrando

AGRANDO ZÄHLT ZU DEN FASTEST GROWING STARTUPS Jeden Monat erstellt GlassDollar eine Liste mit den 100 am schnellsten wachsenden Start-ups von Deutschland. Über verschiedene Kriterien wie z.B. Web Analytics und Arbeitnehmer bewertet GlassDollar das Start-up mithilfe einer selbstentwickelten Formel, die dem jungen Unternehmen einen Rang…

Zum Artikel