CEO's are constantly bombarded with new technologies which are always labeled as 'revolutionary,' 'critical,' and a 'must-have' if they want to survive and grow their businesses. The last two years saw the emergence of AI, ML, and other new breeds of technologies. Along with them, the CEOs became acquainted with a new 'species' in the office, the Data Scientist (DS); the very expensive new employee who holds, as they say, the future of the company in their hands and identifies the secrets to its success.
Considering the huge AI and ML challenges and the business benefits they embody, CEOs have learned to appreciate the DS's capabilities and the positive business transformation they can ignite. For example, according to KPMG’s Global CEO Outlook survey, data scientists skills were identified as the most important skills that UK CEOs were looking for to support future business growth. The survey included 150 UK leaders plus 1,150 CEOs from across the world, who were asked about their future investment plans and the challenges and opportunities facing their companies.
KPMG's survey showed that 69% of respondents named data scientists as an important workforce capability in supporting their future growth plans, followed by emerging markets experts (57%), and emerging technology specialists (55%). Looking to the future, 61% of the UK CEOs surveyed said that they are pre-emptively hiring new skills, regardless of future growth targets.
Interestingly, the findings showed that UK business leaders had more confidence and a better attitude towards market conditions and a commitment to seeing through the challenges of the future. In contrast, the CEOs from the rest of the world were less confident as a group with 52% stating that they were waiting to achieve certain growth targets before hiring new skills.
But who will be the champion of the process of developing and implementing extremely complicated AI and ML systems? The data scientist, who is the only one to fully understand the technology and what needs to be done, or the CEO who is the only one who has the power to transform the business truly? The answer is not clear cut.
Of course, the DS must revolutionize the technological landscape and open the company to AI and ML. However, the CEO must drive cultural change to support AI and ML initiatives. This includes much more than words or mission statements — the commitment must be shown in action that points the company in the right direction. The C-suite must move away from instinctual, gut-feel decisions, and learn to trust what the data reveals and act on those insights. The CEO, with oversight of the rest of the C-suite, must be the one to own the overall big data strategy and set the agenda for it. This will enable each business unit to enjoy the benefits of big data.
So, what can the CEO do as the AI and ML champion?
Both CEOs and DSs need each other to succeed, and they must find common ground to cooperate and understand each other’s problems and requirements. It is the job of the DS to lead the CEO to the promised land, to help the CEO understand AI and ML, their true significance and potential effect, and how to use them efficiently. The CEO must learn AI and ML processes and terminology and how to transform a business goal into an AI initiative.
For example, the most crucial aspect of building a ML model for business predictions is problem definition. Do this badly, and the data is worthless. Only well-defined problems can efficiently serve the business goals of the CEO. The DS understands the intricacies and variations of a problem definition, but the CEO doesn’t, nor the importance of the process involved in accurately defining the problem. This is the point where the relationship between the CEO and the DS pivots. Where before the CEO championed the DS, now the DS provides mentorship to the CEO, explaining the concept of problem definition, its role in developing a model, and the benefits one can expect. The DS should provide real-life scenarios that can determine the future of the business
For example, how to tackle the challenge of predicting 'Churn' and formulate an impactful and accurate problem definition? For this purpose, the CEO must participate in implementing the three phases of the ETA model: Entity, Target, and Attributes.
The Entity defines the "subject" of the model. It should answer the question: Who or what is the prediction about? For example, if the CEO wants to predict which customers are likely to churn next month, then the Entity is the customer. To identify customers, you need to select their unique IDs.
The Target identifies the outcome that the model should learn to predict. It should answer the question: What eventually happened to the Entity. For example, if the CEO wants to predict which of the customers are likely to churn next month, then the Target is the churn history of customers.
The Attributes are the various variables that the model will consider to predict churn. Attributes usually answer the question: Would this data might be useful in predicting churn. For example, if the CEO wants to predict if a customer is about to churn next month, then the Attributes would include the various properties that define customers, such as age, gender, etc.
The parameters are connected to the relevant source data, and the model is ready for training.
Once the CEO dives into this process and fully understand it, she can engage in a fruitful dialogue with the DS, and together they can calibrate and refine the definition of the problem, and from this expect to receive qualitative business predictions that will decide the fate of the business.
A similar process can be applied to a wide array of business predictions such as:
Nothing in the AI and ML domains will ever happen without the leadership and active support of the CEO. However, nothing will happen unless the Data Scientist provides mentorship to the CEO introducing her to the intricates of this new lucrative but challenging world.