The first relates to job losses. This is simply not true, although they must have heard the opposite. MI does not replace man, but frees him to deal with more interesting and valuable tasks. Many people around the world are doing work that is far from exploiting their skills and potential – and they are even bored to death. Artificial intelligence is not bored, adventurous, and not mistaken. Whatever the task, we can perfectly entrust tasks I don’t like. The live employee checks the results and uses the information extracted from them in the course of meaningful work.
Think of this as follows. The work of security guards or police officers who review cable TV recordings is effective if they succeed in identifying a crime. Hours spent reviewing uneventful recordings cannot be considered a useful pastime.
Slow value creation?
The other preconception states that MI is extremely complex. Legend has it that it will take two years for the first results to show, and there is a high risk that MI will not live up to its promise. That is, a mere waste of money.
If this was once the case, it is certainly not the case today. There’s no question that a development team would have to robotize in a dark room for years to succeed. With modern, agile co-creation processes, end users can quickly provide feedback and get useful output at every step of the process. Thanks to a gradual, results-based approach, the team gains more and more confidence as the development stages progress.
The first step is a ‘hack’ where a small group shows what can be achieved in a given area in a few under the Sun. This is a low-risk experiment – if there is no result, only minimal investment has been made. And if there is, that is great news. This is followed by the Proof of Value (POV), in which the basic ideas are developed and then tested according to pre-defined success criteria.
If things are still promising (and if not, the investment risk is still very low ) may be followed by a prototype, the Minimum Viable Product (MVP). The point of this is that the first schematic version of the solution is created based on the POV, which shows the whole process and also how it will work. In the next section, we enter the hypercare process. A small group of users then tests the prototype and shares their experience before a rollout, which encourages wider use.
The agile scrum method also breaks down these steps into two-week sprints to minimize risk. in order to. They provide two-week customer demos to ensure that the development meets expectations. In addition, all parts of the project are assigned directly to business purposes for real value creation.
My situation is unique
If they are not worried about job losses and slow, expensive development, many will be held back by the awareness that “artificial intelligence is certainly only for others”. One’s own problem is always too unique, too difficult to solve. Developers will simply not understand. MI is only suitable for companies like Facebook and Google.
Let’s take a closer look at what is a legend and what is a reality.
In that it is a unique problem, it may be truth. Many jobs and tasks seem unreasonable because they take too long or are too labor-intensive to be done economically. This is valid as long as they are not handled by MI. Let’s look at how breakthrough artificial intelligence has brought in the scoring processes of World Gymnastics Championships, or how it supports radiologists in recognizing brain aneurysms on CT scans.
At Fujitsu, we are neither gymnasts nor radiologists. It is true, then, that our technology professionals may not fully understand the challenges facing the client, just as you do not fully understand how to solve them with MI. But if you combine your team’s knowledge of the challenge with our MI expertise and co-creation approach, a solution can be found quickly.
How? We connect those responsible for business challenges – those who understand and know the problem better than anyone else – with our MI professionals. The joint team can already address the issues, question the suggestions, and provide feedback as the solution is developed. Pointing in direction in development ensures that the solution really meets the requirements. Together, those responsible for business challenges, end users, developers and data scientists are able to create truly relevant MI solutions.
We solve the impossible
They may face a challenge that could lead to far-reaching results. At the same time, they do not dare to cut in because the task seems unmanageable, uneconomical due to the cost of education and labor, or simply insurmountable.
Such challenges are so enormous that sometimes one does not even think they can be overcome. But artificial intelligence changes that.
This is not just a theory. It is worth taking a closer look at the R&D co-creation project between Network Rail and Fujitsu MI. We are developing an MI tool that automatically identifies, catalogs and visually displays rail devices and their exact location. The case study shows how Fujitsu has put the above principles into practice and how it has created tangible business value.
