The e-commerce sector is less concerned about the risks of using AI. Half of the e-commerce companies do not see them at all, and the other half are concerned exclusively with cybersecurity issues. They see the key problem on the way to implementing the technology as attracting specialists with the necessary profile, as well as the lack of AI-based solutions and tools on the market or problems with their development.
Among the difficulties panama whatsapp number database that retailers face, Vladimir Yakimkov particularly noted technical integration with existing systems, a shortage of qualified specialists and the need for significant initial investments. Of the most popular risks, he singled out those related to privacy and data security, and in order to minimize them, in his opinion, it is important to use modern encryption methods, strictly control access to data and conduct regular security audits.
Inna Papikyan believes that the main business risk is related to the unpredictability of the development of the situation in a company that has decided to implement AI: "If you rely on AI as a solution to all sales problems, you may find yourself in a situation where the company no longer understands the buyer. It is always important to clearly understand why user data is collected and how applicable the metrics are for their segmentation at the current moment. Buyers' needs may change, and AI will not have time to process enough data to identify a new trend. In any case, all risks can and should be foreseen before the active implementation of AI in the company's processes."
Ilya Aboimov named the main risk of implementing artificial intelligence as its instability - the algorithms used can have hallucinations and periodically produce unexpected results: "To avoid this, the same services in the e-commerce sphere, offering services based on artificial intelligence, add the functionality of confirming the proposed options for improving product cards or their SEO - without confirmation from users, changes are not made. Therefore, it is not yet possible to fully trust the functionality of artificial intelligence algorithms. It is expected that in the coming years the problem of hallucinations will be solved, which will allow for the automation of process optimization."
Sergey Beloklokov sees the main risk in the high cost of AI implementation projects: "At the same time, projects do not always take off as the company would like. The result is a loss of time, money and, perhaps, even demotivation of the team for any subsequent similar decisions. There is also a difficulty in that it is not always possible to find competent personnel to implement such tools. And for this, you have to cultivate these personnel within the company, and this takes time and money."
Dmitry Chuvikov sees the main difficulty in implementing AI in practice in collecting, processing and analyzing data: "In order for machine learning algorithms to work effectively, they require high-quality and diverse data. In online trading, this can be especially difficult due to large volumes of data, its diversity and the need for real-time processing. In addition, it is important to consider privacy and data security issues when collecting and using user information."