What managers need to know about AI / artificial intelligence?

While most companies have introduced various roles that interface between the business and the IT, comparable AI-related roles are frequently missing. Indeed, having “analytics translators” – individuals that link day-to-day business operations to the data and AI team – is one of the most important factors for bringing data and AI successfully into action. Organisational goals are usually not directly related to data and AI. They rather circulate upon hands-on topics such as increasing revenue and profits or generating more attractive leads than which algorithm should be chosen for a specific problem. As a consequence, data scientists frequently lack the required business understanding to directly provide value. This is exactly where analytics translators step in. They take up these goals and translate them into data-driven questions and use cases that can be solved with an AI toolbox by the data and AI teams. At the same time, they facilitate the acceptance of the developed solutions within the business - without building trust in the results, AI is not likely to provide any value.

But what do managers need to know for becoming effective analytics translators? For building our executive program CAS HSG Big Data & AI for Managers we intensively investigated where successful analytics translators must excel for playing a key role in the data-driven transformation of their organizations.

Here are our results.

"We never through away data" - This quote from Jeff Bezos illustrates the central role of data as fuel for Amazon's business model. However, even at Amazon, data alone does not create competitive advantages. Data can only create value when it is used, e.g., to improve decision-making processes or develop new business solutions. Managers must understand the rules of a digital and data-driven world and how their products, services, or processes must be adapted. This also involves developing an AI strategy and identifying, evaluating, and implementing potential uses cases by means of agile innovation processes (design thinking, prototyping, etc.).

80% of the work in AI projects involves data engineering –processing, preparing, and cleaning data. Managers need to be familiar with basic organisational and technical issues of data engineering to assess the effort and feasibility of specific AI projects and how data can be effectively re-used for other contexts once it has been collected. This includes not only an overview of common analytics infrastructures and tools but also an understanding of data management including potential legal and ethical issues that may arise when handling data.

Even if analytics translators do not develop AI models on their own, they must know the language and jargon of Data Scientists and the basic steps of model training and evaluation. They must be able to judge chosen AI approaches from a business perspective and to explain them to other stakeholders within the organisation. Many companies are therefore starting to train their employees for this role because it requires a deep understanding of the organisation’s products, processes, and decision-making structures.

While many companies have successfully implemented initial pilot projects, many fail to scale them. Managers need to design organizational structures and processes with which they can deploy and integrate their prototypes into regular business operations. Companies, such as ZF Friedrichshafen, Volkswagen or Siemens, have set up "AI Labs" in which project ideas are turned into prototypes in the shortest possible time. However, building data-driven organisations is not just about building scalable prototyping factories. Rather, it often involves the systematic redesign of decision-making and business processes and accompanying change management.

Successfully implementing AI in organisations, requires effective collaboration between data and AI teams and the business. As analytics translators, managers can occupy a key position here and play a leading role in the data-driven transformation of their companies.