Jan 24, 2022
6 Things to Know Before Implementing AI in Your Business
Artificial intelligence
If most people have heard about artificial intelligence and the many benefits it can bring to our business models, few of them actually know how to implant it in their company, and many wonder what steps to take before embarking on an AI project.
Alexis Fortin-Coté holds a bachelor's degree in mechanical engineering, a master's degree and a doctorate in robotics, and completed postdoctoral studies in data analysis. Now the lead AI scientist at Vooban, he came to answer 6 questions about artificial intelligence that we regularly receive from our clients.
1. In your own words, how would you explain artificial intelligence?
“Artificial intelligence is a computer system that can make decisions according to data presented to it, just like a human would. Its power also lies in its ability to perform ‘superhuman’ analytical tasks, which go beyond our capacity to do calculations.
Instead of using pre-established rules to operate, AI learns by example and creates its own rules based on the data previously received. Because it can perform significantly better than humans, AI is found in a plethora of businesses and industries to carry out a wide variety of tasks.
For example, artificial intelligence is used in computer vision systems (facial recognition, automated driving, optical sorting), in prediction systems (sales prediction, fraud detection, risk analysis) and in natural language processing (vocal assistants, chatbots, message filtering).”
2. What’s the difference between traditional statistics and artificial intelligence?
“Traditional statistics are easily interpretable. People usually know the mean and standard deviation and can apply them to make analysis. They’re still relevant in sales analysis, for example, since they give clear results and are easy to understand and explain.
Artificial intelligence comes into play when we’re faced with more complex problems that require further decision-making and analysis. Some people see AI as a form of ‘advanced statistics.’ They’re not wrong, but there is still a difference in usage between the two. Traditional statistics are used to demonstrate historical facts via linear models, while AI is able to explain more elaborated relationships between variables and expected outcomes. It can use much more data, often disparate or unstructured. It’s substantial work, which goes beyond the limits of human analysis. The more variables and complexity there are, the more justified the use of artificial intelligence is.”
3. How much time should be invested in data preparation before training an algorithm?
“Data preparation is the most crucial and time-consuming step, and often the most underestimated. On average, 80% of data scientists’ time is dedicated to this task. This process demands time and patience, but is fundamental to train reliable models. We’re talking about artificial ‘intelligence’, of course, but we still need something substantial and accurate to fuel it. Many factors can lead to erroneous results, such as missing or incomplete data, uncorrelated batches, typos, etc. This data must therefore be cleaned, that is, prepared and filtered to ensure its accuracy, relevance, and format consistency.
To create a reliable algorithm, we need a data pipeline: a computer architecture that organizes and transfers data from all the sources of information used by a company (databases, applications, Excel spreadsheets, etc.) so that it can be exploited further on. This pipeline must therefore be robust and consistent, and have as few manual inputs as possible. To err is human, which is why the automation of data entry is the method to prioritize. Ideally, companies should no longer use expansive Excel files to store and manage their databases, since they are neither a reliable source of information nor a stable format.”
4. How much data does artificial intelligence projects require?
“It’s a recurring question, and unfortunately it doesn’t have a final answer. Of course, AI doesn’t invent anything: it only looks into the past to make correlations, propose solutions and predict results. Because every business and every project vary so much, the amount of data required will not always be the same. It’s always better to have more than not enough, though, and it’s never too late to start storing and classifying your company’s data. However, its quality is paramount. If you have a bunch of data, but its traceability and reliability cannot be demonstrated, you may end up with inconclusive results.
The beauty of AI is that it can always be trained and optimized with newer and better data. The earlier you start your process, the more your model will be able to evolve over time and offer interesting results.
Conversely, some artificial intelligence projects do not require a ton of historical data. In computer vision, for example, we can rapidly generate information with cameras. In these types of projects, the images produced must be labelled and categorized – a whole different preparatory work, since it requires human intervention.”
5. What kind of expertise does it take to take on an AI project?
“A freshly graduated master’s student in Artificial Intelligence is sure to excel at making models and proofs of concept. For projects to come to fruition, however, you’ll need the efforts of an entire team. At Vooban for example, our artificial intelligence experts all have different expertise that allows them to adapt to the complexity and specificities of our many projects. Beyond AI, it also takes efficient logistical and technological support, like project managers, data engineers and developers, who will make sure the model is functional and scalable.
Above all, we must not forget the clients. At the end of the day, they’re the ones who are behind the projects and who’ll benefit from the final product. It’s important to keep in contact and involve them so that everyone is on the same page on all stages of the project.”
6. How long does it take to complete an AI project?
“Once again, this is a question with multiple answers. The type of project you’re planning on doing will greatly influence its duration. Furthermore, if your company has enough data to its name (data lake or data warehouse), proofs of concept can be generated quite fast. Otherwise, the data must be collected manually with the help of AI scientists, which will take much longer.
Take, for example, a medium-sized company that possess BI (business intelligence) tools and a good sales history. In this situation, it is possible to get a sales prediction model in just a few weeks. On the other hand, if you wish to automate your entire production line and add specialized equipment like sensors or robotic work cells, then we’re talking about a lengthy process that can take place over a full year, sometimes even two.
Nevertheless, with the right team and enough quality data, it is usually possible to complete an artificial intelligence project within a year.”
