According to Monster.com, the three most in-demand skills are: Machine Learning (ML), Deep Learning and Natural Language Processing (NLP), and this is no coincidence. Indeed, as per the growing interest around Artificial Intelligence and Data Science, many companies are interested in knowing more about these technologies and how their companies can benefit from them.
However, these discussions often stay in the meeting room or at the coffee machine for several reasons. Legitimate concerns are associated with AI and its practical use cases. For example, certain employees can be defiant to these new data-driven processes and adoption will likely fail if you don’t have support from key actors. This apprehension is natural and often comes from simple things such as a lack of knowledge. To ease the pain, at Visium we organize workshops to educate and train your employees and executives, helping to bridge the skills and knowledge gap.
Nevertheless, apprehensions don’t always come from employees. Here, we summarise what are the concerns of most executives when it comes to using AI in their own company.
1) It’s Too Expensive
I won’t lie to you. AI can be expensive, as any other business project, but the goal here is to make you understand that it doesn’t have to be.
Indeed, at Visium we encourage our clients to start with small proof-of-concepts or prototypes to test the potential return on investment, minimize the risk and familiarize all parties to Machine Learning. These projects can then be field-tested, e.g. on one manufacturing line for one type of product, and the savings or added value can be accurately measured.
This way, you don’t have to unlock enormous funds to get started with AI. Most of our clients are actually amazed by the return on investment of these small prototypes but also the speed of development and deployment. In the right hands, these technologies can deliver excellent results quickly and at low costs.
2) I Don’t Have Enough Data
Of course AI is a data-driven technology, and not every company has an IT infrastructure like tech companies to gather and store their data since they exist.
However, a common misconception is that AI and Machine Learning is simply throwing data at algorithms will make the dream come true. This is definitely wrong, as Peter Norvig says:
“More data beats clever algorithms, but better data beats more data”
Indeed, AI shines with the right kind of data, not just any data. It would be unrealistic to expect to use a predictive model for sales if you don’t happen to have collected any data on your customers or your leads.
Therefore, before considering working on an AI project, it is important to brainstorm with knowledgeable technical profiles and employees to understand what kind of data the model will need. It is also important to analyze after a project what would have helped during the data acquisition. At Visium we often advise on the best set of actions that a company can take, in terms of data governance, in order to increase its possibilities.
3) My Business Is Not Suited for AI
This is an illusion that emerges quite often when discussing with executives. Indeed, people think their business model is outside the scope of impact AI and automation are having on the market.
Let’s take a look at the following illustration from McKinsey.
This research from McKinsey published in 2017 concludes that every industry and job can be partially automated, and will be impacted as AI adoption grows. In the future, we can even expect to see these numbers increase as the state-of-the-art keeps exploring new possibilities.
Applying AI to a business challenge or to enhance a product is a difficult task. It requires training to shape our thought process around these new paradigms. More than anything: it requires open-mindedness and passion.
Indeed, the most important skill to create value with data is being able to make links between seemingly unrelated subjects. If you had the chance to read our recent blog article about “The Efficiency of Neural Networks and SVM in Cancer Detection” from late August, you could start thinking on how to adapt the approach to your industry specific problem at hand, whether it is in Manufacturing or Marketing. This is how we, by working with dozens of different projects in different industries, we can have a targeted yet creative and outside-the-box approach that creates considerable value for companies.
As for all technologies, experience is key. All these concerns and hurdles can be overcome by brainstorming and discussing with people who experienced similar challenges.