A discussion with Caroline, a Senior Project Manager and Chief of Staff at Agilytic, provides insights into the world of data projects, from common misconceptions to career development. Caroline, who has a background in audit at KPMG and financial analysis at Swift, discusses her transition into data and how companies can successfully implement data and AI initiatives.

The Starting Point for Data Projects

When beginning a data project, Caroline advises business leaders to start with their business needs, not with the technology. Instead of jumping on the latest AI trends, she recommends identifying major strategic priorities and pain points within the company. For example, a common pain point could be the time spent on low-value administrative tasks like data entry. By focusing on a few specific use cases, a company can determine which data-driven solutions might be most effective.

A crucial first step is to ensure the quality and quantity of data. The data must be complete, well-formatted, and free of errors. Poor data quality, even if thought to be 80% correct, can lead to disappointing results. As the saying goes, "garbage in, garbage out". Sufficient data volume is also necessary to derive meaningful trends, which is often tied to a long data history.

Despite the importance of data volume, data projects are not only for large corporations. A company with a limited number of clients can still have a large volume of data if those clients have many interactions or make many purchases.

Real-World Applications of AI

Caroline provides several examples of how AI can be applied:

• Invoice Management: AI can be trained to read and structure data from incoming invoices, even complex ones with multiple pages or irregular layouts. This can save a significant amount of time, potentially reducing the workload by up to 80% for manual tasks.

• Customer Segmentation: Data analysis can be used to segment customers for marketing purposes or debt collection. By analyzing transactional and behavioral data, a company can differentiate between a customer who is simply forgetful and one who is a potential fraud risk. This allows for a more tailored approach, such as sending a simple reminder to a distracted customer versus taking a more assertive tone with a difficult one.

The Human Element in Data Projects

One of the key takeaways is that AI is not meant to replace humans but rather to assist and enhance their capabilities. AI handles repetitive, low-value tasks, freeing up employees to focus on activities that require human expertise and interaction.

For a project to succeed, it must be seen as an organizational effort, not just a technical one. All stakeholders should be involved from the start, including management, field teams, and technical staff. This ensures the solution aligns with the company's strategic priorities and the realities of day-to-day operations.

Career and Professional Development

Caroline also touches upon professional growth in the data field. She emphasizes the importance of a well-rounded skillset that combines technical knowledge with an understanding of business challenges. She recommends aspiring data professionals not to be afraid to ask questions and to maintain a curious attitude toward business problems. For career development, she suggests a philosophy of "daring" to step outside one's comfort zone, even if it means taking a step back initially.

At Agilytic, new employees undergo an onboarding process that includes foundational training and practical exercises. Projects are typically kept short, around three months, with tangible deliverables at the end. This structure allows consultants to quickly gain exposure to a variety of industries and business problems. The career path progresses from junior roles, where consultants are coached by senior colleagues, to more autonomous senior positions, and eventually to management or expert roles.

Listen to the entire episode (in french) here.