Deploy a Successful AI Strategy // Part 4: Know-How

According to the prestigious Harvard Business Review, being a data scientist is having the “sexiest job” in the 21st century.

This statement from 2012 is certainly true in part, considering how omnipresent the topic of ML and AI has become in our daily lives. Whether it’s Google, Netflix, or Amazon, all are making above-average use of these technologies that have propelled them to become the most valuable companies in the world. The business network LinkedIn has also determined that in the past three years, the demand for data scientists has grown by over 37% p.a.

Know-How anyone? Yes, more please!

The problem: It is difficult to find people with this know-how, to acquire them and to keep them in the company in the long run. The requirements for this job profile are high: In addition to mathematical/statistical skills, good knowledge of programming as well as basic IT infrastructure such as cloud or storage technologies is essential. Furthermore, it is important that a data scientist brings along a certain expertise of the respective industry in order to develop really meaningful models.

In order to equalize this problem, it is advisable — depending on the size of the company — to divide those tasks into three roles:

• Data Engineer: Collecting, preparing and providing data.
• Data Scientist: feature engineering and development of models
• Business Analyst: Collecting, preparing and evaluating requirements from the business departments

Data Science for all!

The topic of continuing education is also an essential component: The dynamics in the field of ML and AI are extremely high, so many company resources should be invested in employee training. From my own experience as an employee and manager, I know that this topic is an essential criterion for retaining IT employees in the long term (this behavior is particularly important among Millenials and Generation-Z).

In the context of continuing education, it is also important for the entire company to gain a basic understanding of working with data (“data literacy”). The more this literacy is embedded in the corporate culture, the easier it will be to siphon off insights from corporate data in the long term. Even if it looks like a “boiling the ocean” endeavor: Try to train all employees in this discipline. Some will have reservations, others will actively resist. And let’s face it: you won’t be able to please everyone. So target the 80% of the workforce and form smaller teams early on to lead by example (more on this topic in the “Culture” chapter) and ensure acceptance over time.

Key Take Aways

• What should be the sourcing relationship of internal/external data scientists and engineers?
• What roles are needed in your own use cases? Is role splitting desired or even required?
• What incentives can be used to retain talent in the long term?#
• What do scalable training concepts look like?
• How can a general “data literacy” be anchored at all levels in the company?

Want to read more? Return to overview