Previous Article: Vooban and Mila Join Forces for the Development of AI in Quebec
Next Article: Artificial Intelligence, an Ally to Overconsumption?

Mar 23, 2022

AI Projects: Where to Begin?

by Hugues Foltz Executive vice-president

Artificial intelligence

I have been involved in Quebec’s artificial intelligence (AI) industry for more than five years. Having almost always worked with new technologies during my career, it was only natural for me to rapidly gain interest in the many possibilities offered by artificial intelligence.

Like many others, I immediately saw the tremendous potential of this technology for industries of all kinds. After reading thoroughly on the subject, I knew that it would forever transform the business world as we knew it.
I believe most of Quebec’s companies have realized that time is running out and that things must get moving. That being said, I understand that the many steps of a digital transformation can appear laborious, and it’s totally normal to seek guidance during this process.
In my last article, I talked about the people you should hire to carry out your company’s AI projects to maximize your chances of achieving the desired results. This time around, I’ll help you get your first AI projects in order and identify where they could have a significant impact!

There are so many use cases that they can certainly become a little disorienting. First things first, you need to look at your organization with brand new eyes. The objective of your project is not to map all your processes in great detail just to optimize microtasks or make small gains here and there.
Instead, the goal is to outright eliminate certain tasks by replacing them with technology and automation! It's an entirely different mentality, which is essential to seize all the opportunities brought by a digital transformation and the automation of your processes. 

Here are some tips to give you some insight and ideas before undergoing a digital transformation with AI. These are often classic wins, applicable to many business models:

  • If your company is still dealing with tons of paperwork, consider digitizing that data.
  • Do you have Excel files with which you manage key processes? If the answer’s yes, you should know that all the data compiled in these files can’t be used to optimize your working methods or launch your first AI projects. We need to stop using Excel as a management software and start storing company data in appropriate databases. Excel is a spreadsheet, nothing more!
  • If you’re still using your eyes to do quality control, it’s high time to let the machine take over! Humans are not good when it comes to continuously look at objects to find flaws and validate dimensions.
  • If humans do the planning of your operations, it is certainly not optimal. When we talk about calculation and analysis of many variables, the machine will always be better, and above all, much faster.

Now, if you're wondering where to begin your first AI projects, just apply these few simple principles. First, start by looking at the places where you make a lot of money – the processes in your business model that are highly profitable.
Next, look for expensive processes that are making you lose money. Both situations are usually great starting points for using technologies like AI to optimize your company. The reason is quite simple: as a general rule, technologies will bring you gains of 3% to 5%. Obviously, projects with such gains are much easier to benefit from if they are applied to very profitable or very expensive processes...

Ask yourself what’s harming your company. What are its weakest links? Amongst your tasks, which ones have become too dependent on humans and would benefit from being automated? Once you’ve answered these questions, you will have a better understanding of your business model and of the places where technologies can have the most impact!
There is one last thing that I would like to clarify with you. Too often, people reach out to me on LinkedIn asking for advice on this very specific question: is it easy to become an AI scientist? The answer is no. One does not simply become an AI expert, at least not without having completed the necessary studies, which means at least a master's degree and ideally doctoral studies in data science or a corresponding field.
The belief that a few online courses on Udemy or Coursera will allow you to become a data scientist worthy of the name is unfortunately an illusion. Be vigilant and take the time to carefully validate the resumes of the so-called AI experts that you hire or that are referred to you.
Basically, what you need to remember is that technologies are advanced enough and they must be seen as an essential ally to the prosperity of your business. The world is changing at an incredible speed, and it is more crucial than ever to initiate a technological shift to ensure the success of your business model for the years to come.
All you have to do is surround yourself with strong, reliable partners and set up an action plan that will harness the full potential of artificial intelligence and new technologies. Be innovative, be bold, and get moving before it's too late. In fact, no business can afford to lose the race against digital transformation!