We are opening a new campus in Munich! Join our Full-Stack and Data Science batches starting in February 2021.

AI (Artificial Intelligence) for SMEs

AI (Artificial Intelligence) for SMEs

AI (Artificial Intelligence) for SMEs

Published on 26.10.2020 by Dr. Nitin Kumar
This blog has been produced to share my experience of working/talking with Small and medium-sized enterprises (SMEs) in Switzerland on the topic of “How to become an AI driven Organization?” Though the topic cannot be compressed into one single blog, questions related to general awareness around becoming a Data-Driven Organization are presented here.
 
SMEs in this article include both companies between 1 – 100 and 101 – 999 employees. This article primarily focuses on AI passives (1): companies that have no Artificial Intelligence adoption and/or little to no understanding of AI.
 
 

The very first questions?


The first and only question AI passive companies should be focusing on is: why Artificial Intelligence? What made them think they need AI? Is it because of the competition getting an edge, management wanting to be the part of the hype, or a deep understanding at the management level on what AI can and can’t do? If you fall into any other group except the last one (management knows why and how AI can help?), you don’t want to start with AI as of yet. I assure you you will fail! If you are at this stage, one option is to have an AI awareness training for the management team (and other decision-makers) on topics around strategizing AI.
 
 

Chief Decision/Data/Analytics Officer


The first and foremost person you should think about having on board to really become AI- driven is a Chief Data/Decision/Analytics Officer. Such a person ideally has a good understanding of both your Business and the potential of AI (on the applied side rather than algorithms).

This individual, or select few, does not need to be hired from the outside (they don’t exist, I assure you that) but rather taken from internal management and ready to learn Data and AI. Without someone who takes the lead across departments, helps people select use-cases, prioritizes and delivers them, the data journey is most likely to fail. One can also start projects  across departments without an AI leader on board. Such an approach might work in the short-term, but if you are really focused on becoming an AI driven organization, think about this title from the beginning.

  

Use-case Identification


Once you are convinced that AI is the way forward (which I believe you would be, if you look closely into any business), the company should focus on the identification of use-cases that are really data-driven. Such focus is very important as most companies tend to get lost in technology and forget about the business. If you are not sure how to identify use-cases, I highly recommend you to do an Analytics Translator training that helps business professionals identify and define them (feel free to reach out to me regarding that). The use-case selection should concentrate  on keeping ROI and implementation costs in mind. Low hanging fruits should be targeted first keeping the long-term goal in mind (ex: becoming Data Driven).


 

Buy vs. Build


This is a question I have been asked by several people: what if we are able to find a pre-built solution for our problem? Should we buy it? The answer to this question depends on what that solution will do for you (basically the ROI) and how long it might take to implement. If the solution can generate high ROI immediately (keeping the integration and recurring costs low) but takes significant time to implement if done internally, then go for it. If the solution generates high ROI and takes less time to implement, you could also implement on your own, either internally or consider an external service provider.

It is always tempting to buy pre-built solutions. If you are deciding to go down this path, I recommend focusing on only one question – should I keep buying technology stack for every data problem I face or should I think about an AI strategy for my company?
 
Important warning – Data keeps changing from time to time. What will happen to such a solution when the data it uses has changed (feel free to Google Model Decay and Model Drift)?
  
 

External vs. internal


This is the second most important question I hear from SMEs. Should we hire someone, or should we go with a service provider?

My answer is the same as for the Buy vs. Build with a special note focusing on the warning above. Since this blog is for AI passive organizations, finding external help in the beginning might not be a bad idea. However, SMEs should really pin down the use-case they want to solve. Once identified, they should follow the approach defined in Buy vs. Build.
 
 

Hiring and Reskilling


If the goal of a SME is really to become AI-driven and have it as part of their future strategy, then they should hire the right people. Who to hire will depend upon the current state of their data. If any data related to any problem they are trying to solve is in silos, then you potentially want to hire a Data Engineer and Data Scientist. The Data Engineer should focus on building a data warehouse (or lake or swamp depending upon how much data you have). The focus here should be keeping some well-defined projects in mind. A company-wide data architecture project with no ROI in mind is doomed to fail (I have seen that happening again and again). If most data related to your projects is readily available, then you do not need to start by hiring a Data Engineer. Only a Data Scientist should do the trick as they should since they should know the basics of Data Engineering.

One big issue often noticed is that most SMEs in Switzerland are in the space of use-case identification and business-case translation (Analytics Translator). Though there’s huge interest in doing AI, the knowledge around what it really is and takes to become AI-driven is lacking both at the board and senior management level. For this group, I recommend AI-awareness training. For those who will be involved in use-case identification and prioritization, Analytics Translator is the right training. Lastly, AI for Managers is the best choice for candidates who plan to build AI teams. 
 
  

Conclusion


AI has huge potential to deliver growth to any sort of business (as much as 10%). Most companies are struggling to deliver impact using AI, not due to shortage of talent but to the  clear lack of strategy at the management level. The points mentioned here are in no way the only recipe to become an AI-driven organization, but merely food for thought. It is easy to be tempted by multitude of service and pre-built solution providers that are happy to help.,When going for one, however, keep the points above in mind. If you are not sure, consult different experts (I am happy to provide you some free guidance). AI is not about the technology; it is about the system approach that organizations must take keeping data, people (human-centric AI) and long-term strategy in mind to really become an AI-driven organization. IT IS AN IDEOLOGY! Last but not least, AI is an unexplored territory for most companies and almost all of them are struggling with it, YOU ARE NOT ALONE!
 

Upcoming Dates

Full-Time Programs, On-Campus or Remote

Apply to one of our Programs

Apply by the early application deadline. Otherwise a late application fee will be applicable.
Accept