AI is on everyone’s lips and can, among other things, support B2B commerce in various areas – but still many AI projects usually fall far short of expectations. In the second part of the series of articles you can read about the stumbling blocks that AI can encounter.
The idea of AI can be found at all hierarchical levels in the company. But on the way from the project idea to the successful implementation of an AI solution some hurdles have to be overcome. These stumbling blocks can develop into considerable risks for a company.
Two significant risks are described below:
The hype of AI projects
There are various “problem groups” for AI projects: projects that can be solved more efficiently without AI or projects that can only be solved very expensively by AI. But also projects that do not generate any real added value or that are technically feasible but not accepted.
AI is not a miracle cure for all problems: Assumptions about the feasibility and benefits are often overestimated – assumptions about the costs are often underestimated. For this reason, most AI projects fail. Companies lose confidence in AI solutions and refrain from re-initiating them.
Due to the success and boom of AI, many tools have been developed that allow many models to be created without expert knowledge. For example, chat bots can be initiated without a great deal of time and effort, which at first cannot be distinguished from a human being.
For larger use cases, however, the effort is more significant: data acquisition, preparation and cleansing. At this point experts with profound knowledge are necessary. Often the system cannot be integrated into the company’s system landscape due to requirements that cannot be implemented, differences in technology or interface problems. The AI project does not fail as such, but in the implementation. Also for subsequent projects no added value can be drawn.
Conclusion: Artificial and human intelligence as a dream team
Even with failed projects in the past, there is no way to discourage. Companies can learn from their mistakes and minimize controllable risks.
With the development of a long-term strategy for the sustainable use of AI, the success of AI projects can be significantly increased. This is done in a harmonious interaction between artificial and human intelligence, with humans setting the strategy and repetitive analyses of complex data sets from machines.