A growing number of large companies are embracing artificial intelligence (AI). Questions arise about internal organization, prioritizing projects and data governance. That said, it's no longer a question of whether or not to adopt AI, but rather how.
AI: Relative intelligence, yet extremely useful
We talk about Artificial Intelligence (AI), yet machines are still a long way from achieving what we mean by intelligence. Take the recent example of ChatGPT, an excellent language model that has no common sense and is therefore not intelligent. Even if AI isn't intelligent, this doesn't prevent it from being useful in a multitude of situations. Today, AI models are used in all sectors: manufacturing, finance, energy and healthcare, for example.
Reserved for research laboratories and large technology companies a few years ago, artificial intelligence is now accessible to all businesses. The factors that have led to this blossoming are manifold: lower storage costs, increased computing power and ease of use of AI tools. That said, for companies to succeed in their data-driven transformation, three complementary axes need to be taken into account:
- use cases,
- talent,
- a common language.
Identify use cases relevant to your business
To identify the most interesting use cases, companies can use a three-step approach:
Step 1 – Identification
One or two brainstorming sessions with stakeholders are an excellent starting point for listing use case ideas. It's important to combine an approach based on needs or desires (top-down) with a reflection on the data available within the company (bottom-up).
Step 2 – Selection
Use cases are placed, roughly speaking, on a benefit/impact matrix. The aim is to separate them into groups. From 2 to 5 use cases are then selected, to form a balance between ideas that are easy to implement straight away, and promising but more complex projects.
Step 3 – Definition
The selected use cases are then detailed using an outline. An example of such a canvas is provided below (Figure 1: the Data Initiative Canvas). The purpose of the canvas is to ask the right questions about the project. It also serves to align all stakeholders with the initiative. Note that the document will evolve over time.
For these three stages, you need to find the right balance between the time you need to devote to them, and the stakeholders you need to involve, if you are to make reasonable progress.
Finding the right talent for your AI project
To develop AI use cases in business, you need to find the right talent or talents. There are several possible options, such as:
- Specializing business experts - Hiring a data scientist is expensive for a small company. You can therefore train a business expert as a data citizen.
- Recruiting data scientists - If the company can afford it, this is certainly the best solution. The development of in-house skills is an important asset for companies in the long term.
- Consulting firms - Outsourcing resources is also possible. An intermediary step is to recruit a junior data scientist profile and seek the advice of a consultant on the strategy to be adopted.
Once this choice has been made, all that's missing is one important ingredient to make the use of AI enterprise-wide.
Speaking the same language
For the successful use of AI in business, it's imperative that employees develop a common language. This is all the more critical for smaller companies, which often can't afford to hire a team of data scientists. To achieve this, we need to create an environment that enables employees to become more autonomous with data.
This is where the concept of data literacy comes into play. Textually, this means knowing how to read and write data. More concretely, it means knowing not only how to interpret data, but also how to communicate it. The idea is to make employees aware of the importance of data, its possible applications and the limits of artificial intelligence.
A continuing education program focused on data management and enhancement
In the field of digital transformation, it is essential that managers develop a solid understanding of the concepts involved in data science. One initiative that meets this need is the Certificate of Advanced Studies (CAS) in Data Science & Management, developed in collaboration by HEC Lausanne and EPFL.
The CAS in Data Science & Management offers a unique program designed specifically for business executives. At the crossroads of technology and management, it focuses on concepts such as :
- Assessing corporate data maturity
- Understanding and implementing an effective data strategy
- Managing data governance and quality
- Mastering data visualization and designing dashboards
- Discovering machine learning and the role of data scientists
To ensure a thorough understanding of the concepts taught, participants carry out a corporate project linked to data management or enhancement.
