The hypes surrounding AI techniques has raised expectations to the levels that short of transforming lead into gold, business expectations are bound to be unfulfilled. The lack of competencies with AI techniques will force organisations to automate the development of AI models, accelerating non-differentiated solutions at the expense of creativity. In addition, data and process complexity will continue to increase rapidly, while business processes remains siloed, intricate, and difficult to harmonize. As a result, AI projects will remain alchemy, run by wizards whose talents will not scale within the organization. That is, until the technology begins to help more with the creative part of the process. Thankfully, research into data science automation is starting to increase faster than data complexity, allowing skills to begin to catch up. For example, automated processes for the selection of AI algorithms are simplifying technical processes, allowing AI users to focus on business problem-solving. The last five years have seen a proliferation of projects and organizations worldwide facilitated by AI techniques and in particular, machine learning and deep learning. This acceleration has been boosted by the adoption of open source languages such as Python, promoting its power and flexibility at the expense of traditional disciplined approaches favoring the readability and homogeneity of analytical processes. The hype surrounding AI techniques has become deafening, generating unreasonable expectations. From a technical perspective, IT systems aren't always ready to systemically integrate smart components, and from a business perspective, AI techniques have started to disrupt marketplaces, including labor marketplaces, delivering value in this technological cacophony, where adoption is still outpacing the production of competent AI professionals, often requires a special type of wizardry to harness and channel the power of AI techniques. This is leading to competitive differentiation. Emerging capabilities such as Auto ML or automated machine learning, the automated process of features and algorithms selection, tuning, iterative modelling, and machine learning models assessment, all promised to bring more science into this artful domain. But they only address a limited non-creative portion of the entire process. Therefore, not necessarily scaling the needed talent. AI techniques not only require the technical expertise of mathematically savvy data scientists, inventive data engineers, rigorous operation, research professionals, and shrewd logisticians, they also depend on the complementary skills of specialists in the adjacent fields such as linguists, ergonomics, and designers. Yet the success of AI projects hinge on another additional and crucial ingredient, tight collaboration with the business through open-minded domain experts. A common language across all of these competencies has yet to be defined. We talked about data literacy earlier. As a result, imperfect translation efforts remain a key barrier for the scalability of this collective talent. In truth, a majority of existing AI practitioners are skilled at cooking up a few ingredients, but very few are competent enough to master several recipes, let alone invent new dishes.