<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Nailya Zamashkina on Medium]]></title>
        <description><![CDATA[Stories by Nailya Zamashkina on Medium]]></description>
        <link>https://medium.com/@NZamashkina?source=rss-4129799654de------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*FGe_RPqupHAVbWigC-4BcQ@2x.jpeg</url>
            <title>Stories by Nailya Zamashkina on Medium</title>
            <link>https://medium.com/@NZamashkina?source=rss-4129799654de------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 05:10:18 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@NZamashkina/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What is artificial and what is usual: can neural networks replace people in the banks?]]></title>
            <link>https://medium.com/@NZamashkina/what-is-artificial-and-what-is-usual-can-neural-networks-replace-people-in-the-banks-ff9275803485?source=rss-4129799654de------2</link>
            <guid isPermaLink="false">https://medium.com/p/ff9275803485</guid>
            <category><![CDATA[startups-in-russia]]></category>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[fintech-startups]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Nailya Zamashkina]]></dc:creator>
            <pubDate>Wed, 24 Oct 2018 19:43:17 GMT</pubDate>
            <atom:updated>2018-10-24T19:43:17.676Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*D4Mnw8eatgR1_q_B8EJpsg.jpeg" /></figure><p>The concepts of “artificial intelligence”, “big data”, “machine learning” have become mainstream in the financial market very recently. However, the transition from theory to a practical implementation of technologies with elements of artificial intelligence (AI) was very fast — today they are used by most Russian banks.</p><p>We should not that there is still a confusion in terminology, even at specialized events. The concept of “artificial intelligence” is not always used appropriately. This is the reason Stanford University almost banned the use of the term. Experts explain that AI does not yet exist, and what is called AI today are actually programs that can analyze data with varying degrees of independence. In other words, we are talking about machine learning, robots, and chatbots with different interfaces.</p><p>The first direction that has started to grow rapidly is text-based chatbots for instant messengers. The development teams of banks relied on a young audience that “keeps holding smartphones in their hands”. However, these were nothing more than pilot projects, the aim of which, in professional terms, was to increase the volume of retrieval and to digitize the so-called “long-tail answers”. Some large banks have achieved impressive results, “closing” up to 70% of all customer requests with automatic responses. The next stage was the rapid development of the market of bots that support the technology of Natural Language Understanding, that is, the understanding of natural language. According to Just AI, by the end of 2018, the voice bot market will double, exceeding half a billion rubles, and in five years it will reach 33 billion. In the West, voice bots are known as”Conversational Commerce “ (the term was coined in Uber in 2015). The main advantages of this technology for the client are individual service, speed, and 24/7 format. For the Bank, it is, first of all, a significant reduction in costs, including the maintenance of call-centers and the speed of response to customer requests. Today, voice bots are the main direction of the AI, including smart speakers and all sorts of intelligent assistants.</p><p>Only 0.5% of customers talking to a robot by Dasha service.AI realize they have been talking to a robot.</p><p>Russia has well-established its own developments in the field of voice biometrics, which use neural networks to determine the most likely phonemes of the Russian language. What is important, these acoustic models of the Russian language are self-learning and able to adapt to the regional dialect. For example, only 0.5% of customers talking to a robot by Dasha service.AI realize they have been talking to a robot.</p><p>A separate direction of using voice assistants is the so-called soft collection. There have already been the first projects to use robots for calling customers with overdue debts. The robot is still a little inferior in these communications man, but it has an undeniable advantage: the robot will not break the rules (i.e. the conversation script), it will give in to emotions. This is extremely important when dealing with clients — regardless of whether they are individuals or legal entities. Anyone can be spoken to emotionally.</p><p>This summer, the Unified Biometric System (UBS) was launched in Russia. It allows to remotely conduct banking operations. But new opportunities also carry serious risks: social engineering is an integral part of fintech, which means that we are already dealing with an almost endless shell and armor competition. All existing methods of identification are imperfect and vulnerable — there are developments that imitate voice and image in real time. The solution is not only in an integrated multimodal approach but also in new AI technologies that can ensure the safety of data exchange.</p><p>Robots read through dialogs between the operator and the customer and dialogs fall under quality control according to specified criteria.</p><p>AI has been successfully used for controlling functions within a bank. For example, robots read through dialogs between the operator and the customer and dialogs fall under quality control according to specified criteria. The scope of AI extends to the interaction of banks with the regulator in the field of compliance and KYC. And, of course, the improvement of the scoring systems is a revolutionary direction for the bank — there is no way to cope without AI elements, as the priority is given to social scoring, and not to allow commercial services to have unlimited access to people’s daily operations.</p><p>Banks actively use neural networks as well. And there are several directions. One of the promising areas is the preparation of claims to the court on the basis of the neural network. Increasingly, they are used to predict borrower behavior in the future: the network is self-learning, and the more data is loaded into it, the better. Finally, AI has been successfully used in various banking consulting services, from the assistant in the selection of exchange tools (robo-advisors) to the personal budget maintenance assistants. SimpleInvest and TheWaay projects confirm the high demand for services that facilitate the customer’s choice. If you would like to save time, recommendation services are what you need.</p><p>The examples given above are not futurology, but examples of the use of AI at the level currently embedded in banking processes. It is important that this subject got rid of unnecessary hype, and bank strategists decided on the future — AI is actively introduced into many business processes in banks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ff9275803485" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>