However, as seen previously, the main problem for organizations was not about data management, but rather about defining a clear business strategy. Obviously, the tool should answer a real need and show a positive business outcome. Hence, the realization of any Artificial Intelligence model should start with a real business question that needs to be answered cite{mgmt_make_analytics_work}, instead of directly trying to use the data. There exists many innovation and prototyping tools such as Design Thinking which can help in constructing such project.

Design Thinking is a creative tool used for innovation based on building empathy with a user by observing and understanding his needs and problems cite{mgmt_d_thinking}. This method implies spending a lot of time in the observing phase. The way the user thinks, feels, does and says has to be analyzed to be able to learn insights before prototyping a solution. The prototypes should then be tested on the field to obtain feedback that will allow to iterate. In the case of tools to predict the risks of fires in buildings, data scientists and top-level managers should observe and understand the work of the inspectors, their criteria and expertise, before building a predictive engine. Adopting innovation tools will help the development of coherent Data Science projects.Furthemore, Tom Fawcett suggested to decompose every business problem into subproblems before establishing relationships with small Data Science projects cite{mgmt_data_driven_dec}. In fact, a good way to test the business value of a Data Science project would be to start with small achievements and iterate several times until a complete framework can be deployed.

This strategy was adopted in Amsterdam, when the city developed several projects based on Big Data cite{mgmt_mit_amsterdam}. Ger Baron, who is the city’s chief technology officer explained in this article published in the MIT Sloan Management Review that the city launched many projects as small experiments in areas where it was easy to iterate fast before deploying them on the ground. For instance, he developed a project to reduce the public lighting if there were no people around. This project involved many actors such as the company managing the electricity, the city’s department of infrastructure and of course the department of lighting. To facilitate the implementation of the project, he decomposed it into small pilots projects that were successfully deployed later. This approach would help to decrease the uncertainty that exists about the times and investment needed, as noted in the study by The Boston Consulting Group cite{mgmt_bcg_mit}. It would also help to iterate faster and to learn from experience, and it should be followed by organizations who are adopting Data Science. To overcome the cognitive biases, Data Science and its algorithms, which are supposed to be neutral in processing the information, will help, if data scientists can create tools that increase the rationality.

More precisely, the tools created by data scientists should be able to present the right dose of information and be combined with the users’ human-judgment and expertise cite{mgmt_mit_all}. But still, managers may have difficulties with adopting the right mindset to use Data Science tools. Fortunately, making them better at understanding and using information is possible thanks to training and coaching. Several authors in the literature suggested using workshops cite{mgmt_data_decision, mgmt_mit_all} where data specialists could help managers to refresh their statistics knowledge and show how to work on small problems of Data Science using data from their own company. They also advocated for regular coaching and training by experts so that managers could directly ask questions and obtain a personalized help cite{mgmt_data_decision, mgmt_mit_all, mgmt_kotter}. In the research report by The Boston Consulting Group, the approach to build skills related to data science appeared to be different depending on the clusters of companies cite{mgmt_bcg_mit}. The pioneers relied more on hiring and building an internal expertise, whereas the investigators were outsourcing the development of the predictive tools. These observations were coincident with the way Firebird (outsourcing) and Firecast (internal expertise) were built.

But the authors made similar recommendations and advocated for managers to develop an intuitive understanding of Data Science, by learning the basics with online courses and by doing basic exercises. When implementing a big change in an organization, the experts of Organizational Change Management emphasize on implementing and communicating a vision. For instance, John Kotter, who is a teacher at the Harvard Business School, insists on developing a philosophy to show where the organization is going in his famous book Leading Changes cite{mgmt_kotter}.

By developing a short summary of values and creating a picture of the future, it helps the managers to adopt this change. Obviously, the scenario must be feasible and appealing so that employees can embrace this vision. Then, a few advocates that agree with this vision and are ready to adopt the change should be found. Indeed, these persons can help the vision to be spread in the organization by building a coalition that support the change. This group should be composed of senior executives with experience in various field such as marketing, human resources or technology. They will bring credibility and trust to the new process when communicating this vision to other employees and managers.

Finally, an action plan has to be defined with small objectives that can contribute to realizing the overall big picture. This reminds the process of splitting the project into several small tasks, as presented before with the case of Amsterdam. In the case of smart firefighting, the organization should define a clear picture like “using data to develop smart firefighting and improve our inspections”. A group of experts composed of experienced chiefs in buildings inspection planning, experts in Information Technologies and Data Science(who can come from an internal or external structure) and firefighters with experience of visiting buildings should work on the project before showing the improvement that can bring a predictive tool on this task to the fire department.


I'm Erica!

Would you like to get a custom essay? How about receiving a customized one?

Check it out