Page 117 - Azerbaijan State University of Economics
P. 117

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.77, # 1, 2020, pp. 113-132



                    Table 1: Three categories of AI













                    Source: Gonfalonieri, 2019

                    In  the  first  case,  the  AI  solution  focuses  on  improving  and  optimizing  current
                    processes. These usually involve significant development and operational costs. In
                    the second case, which is similar to SaaS, we are talking about interacting with other
                    systems,  such  as  CMR  or  ERP  systems  associating  AI.  The  third  case  is  the
                    expansion of existing AI tools within the company. In every case, the existence of
                    data is very important, but the companies do not need all the data, as they only have
                    to count on the ones that answer the question to the problems that exist. On the cost
                    side,  when  choosing  an  AI  model,  the  costs  must  already  be  integrated  into  the
                    business model (Gonfalonieri, 2019).

                    Once the work teams have been set up, AI integration should begin in small steps, be
                    gradually rolled out to verify that the AI  represents the expected value, and then,
                    after verification and feedback, be continually extended to AI technology across the
                    enterprise. It must be ensured that the new technology created does not replace the
                    daily  routine,  but  complements  it.  Keep  in  mind  that  employees  may  think  AI  is
                    taking their jobs. Therefore, it is necessary to ensure that their scope and activities
                    are expanded so that the AI complements and facilitates the execution of the tasks
                    (Marvin and Horowitz, 2018).

                    According  to  Andrew  Ng  (2018),  who  is  the  chairman  of  Landing  AI,  for  a
                    successful implementation, a pilot project must first be implemented within an easy-
                    to-implement project. Planning and merging teams is the second step, namely, the
                    collection of an expert team but with the help of external experts, of course. The
                    third step is to organize a comprehensive and detailed AI training for those affected.
                    As data collection and appropriate structure design are essential, the development of
                    an  AI  strategy  is  the  next  phase.  Developing  both  internal  and  external
                    communication is essential for implementing a business model and strategy and for
                    proper interactions and cooperation.


                                                           117
   112   113   114   115   116   117   118   119   120   121   122