Financial sector in race for the artificial intelligence

Cutting edge technology is revolutionizing the rules the financial sector has operated under so far. The institutions that will only evolve instead of undergoing a revolution could disappear from the market.
Financial sector in race for the artificial intelligence

(, CC BY)

Artificial intelligence (AI) is entering today’s business like the steam engine entered 18th century industry — rapidly, on a wide scale, and with the promise of irreversible change. According to the latest report of the US computer publisher IDG, global expenditure on AI and cognitive systems will already reach USD19bn this year, over half more than in 2017. In 2021 this figure will reach USD52.2bn, while three quarters of all business applications will be based on AI. Banking accounts for the largest part of this expenditure (20 per cent), closely followed by the medical industry and retail sales. This is not surprising, because advanced data analysis, computer vision, natural language processing, and machine learning, all of which make up cognitive systems, are of special importance to these sectors.

Streamlining processes

In the case of banking, these technologies have been used for many years. The first neural networks to automatically take credit decisions were created over 20 years ago. Banks are also already using on a mass scale so-called data mining or text mining, in other words, extraction of entities from documents or from clients’ correspondence, conversation systems, and analytical models based on machine learning for automation of repetitive manual activities based on permanent rules in such areas as risk, sales and marketing, and counteracting money laundering and terrorism financing.

For example, the British firm Intelligent Voice offers financial institutions transcription tools based on machine learning, which serve to monitor telephone conversations of salespeople as regards the use of confidential information in sales. In turn, Xcelerit and Kinetica have in their offer the tracking of exposure to risk in almost real time, enabling the constant monitoring of the fulfilment of capital requirements. Machine learning is an excellent tool for the automation of financial decisions, serving to assess creditworthiness or the fulfilment of the requirements for granting an insurance policy. This is because it makes it possible to sift through enormous amounts of data.

According to the majority of the 800 managers from the banking and IT sectors who took part in the Accenture Technology Vision survey, the potential of artificial intelligence is greatest in three areas: building customer confidence (71 per cent), optimisation of costs and operations (63 per cent), and improvement of the level of compliance with regulations (62 per cent). Of course, confidence is only possible when the processed data are reliable, verified precisely, and used in an ethical way. 78 per cent of those surveyed by Accenture admit that automation creates new risks, such as false data or external manipulation of data. Automatic decision-making based solely on algorithms also creates the risk of errors to the disadvantage of honest clients. For example, it may prevent making a transfer when the system wrongly “recognises” that it is associated with the risk of money laundering.

Currently, AI is evolving in the direction of offering a broad range of cognitive skills, including feeling, understanding, taking action and learning. These are skills that enable machines equipped with artificial intelligence to interact with humans in a natural way. AI allows to understand not only what the client expects and what he says, but also to identify what he doesn’t know yet. This has its application in, for example, bots (virtual agents), which increasingly replace staff in customer service. China Merchant Bank uses bots in the popular application WeChat to service as many as 2 million queries every day. If not for the application of artificial intelligence, the bank would have to employ 7,000 extra employees.

AI techniques are therefore used for everything — from work in the back office to increasing the effectiveness of customer service. After all, the clients themselves expect modern forms of contact with financial institutions. According to the report Accenture Banking Technology Vision 2018, 45 per cent of clients using digital solutions would like their bank to introduce new forms of communication with them — via virtual reality or the internet of things (IoT). The Korean Hana Bank has met such needs. By using a mobile application processing their data, its clients can obtain the necessary information about the selected apartment or house along with an offer of a mortgage immediately after pointing the camera of their smartphone at the selected real estate.

According to the KPMG report “Rise of the robots”, this trend will mean that in the next 15 years 45 per cent, and perhaps as much as 75 per cent, of the work in the financial sector will be done by robots. So-called robotic automation of business processes will translate into huge savings and streamlining of operations. It is not surprising that from 2019 certified financial analysts will have to demonstrate knowledge of AI in exams.

Financial advice for High Net Worth Individuals is the area in which financial analysis with the use of AI is implemented most rapidly. Robo-advisory systems are responsible for, among others, customer wealth management, financial planning, and creating portfolios of assets. Moreover, by investing in ETF funds (Exchange-Traded Funds), they are to ensure diversification of the portfolio in a relatively easy way with low costs.

One of the pioneers in robo-advisory usage was Bank of America in 2016, whose intelligent virtual assistant uses predictive analysis (data-mining) and cognitive methods to help over 45 million customers make investment decisions. Similar systems are used by such institutions as UBS, JPMorgan Chase, Betterment, and Merrill Edge. According to Business Insider Intelligence, globally, assets valued at over USD70bn are controlled by robo-advisory technology, and by 2020 automatic advisors will control as much as 10 per cent of global assets.

Impact of AI on the economy and labor market

These figures may raise the concerns of customer service employees. According to the research of the World Economic Forum, as a result of the 4.0 industrial revolution (caused by new technologies) by 2021 over 7 million jobs will be lost. According to OECD data, the greatest threat to jobs is in Slovakia, Slovenia, Greece, Switzerland, Spain and Poland — approximately 50 per cent. At the same time, the technological revolution should help create completely new jobs, while 65 per cent of children beginning their education today will graduate with diplomas of specializations that have not yet been invented.

However, it already has a significant impact on global economic growth. During the conference Impact ’18, Microsoft Poland presented the report “Artificial Intelligence of the Polish Economy”, which shows that over the next decade the accumulated impact of artificial intelligence on the global economy could amount to USD1.5-3 trillion, which translates into faster GDP growth by as much as 1-2 per cent.

The authors of the Konzept report by Deutsche Bank Research “Automation — not a job killer”, argue that it is unjustified to blame automation for the global fall in employment of mid-level specialists and their earnings observed in the years 1980-2015. Instead, it was caused by completely different factors (among others, China’s entry into the global market, the fall of the Soviet Union, and global offshoring). Deutsche Bank Research’s report quotes Matt Rognlie, economist of Northwestern, who argues that while technology undoubtedly played an important role in the recent economic growth (citing the USA as an example), it couldn’t have an impact on the share of labor in production since its share in the value of American share capital is still marginal.

Slower development of financial innovation in Poland

In the case of the Polish economy, artificial intelligence accounted for only 0.1-0.2 percentage points of growth in recent years, while only 4 per cent of Polish firms make use of the most efficient mechanisms offered by AI — mainly based on cloud computing. The benefits from the use of AI by Polish companies can be estimated at PLN10-20bn annually, but no more than 1 per cent of GDP.

As Microsoft highlights in its report, despite universal access to the internet in Poland, Polish firms only use the potential of ICT and AI to a limited extent. In each of the areas studied by Microsoft, the use of AI by Polish firms is worse than in the case of their EU competitors. According to the data of the venture capital fund Asgard, in 2017, in Poland only nine start-up hubs worked on AI solutions. In the United Kingdom there were over 120. Over half of all AI solutions are created in British, German and French firms. Those that are created in Poland are often quickly transferred to the West anyway, such as WealthArc, which mainly operates in Switzerland, implementing platforms for managing investments in such institutions as Credit Suisse and Pictet.

“Firstly, failure to adapt or too slow adaptation of regulations to technological change. For example, the act on trading in financial instruments provides that the final financial decisions must be taken by a human, thus excluding the use of robo-advisory solutions in Poland. The attitude of the Polish Financial Supervision Authority (KNF) to innovative solutions and new players on the market is also important. If they do not get the green light from the KNF, they are afraid to implement new solutions in fear of being fined,” stresses Aneta Hryckiewicz, professor of the Leon Koźmiński University and coordinator of the studies program “Big Data in Finance”.

The aim of the KNF is above all to ensure the stability, security and transparency of the market, and not to facilitate the operation of fintech firms. KNF recommendations do not constitute legally binding acts of law, but in practice they are extremely important for the market players and are respected by them. That was the case with so-called screen scraping (which consists in the bank’s clients making their login data available to another bank or entity). The KNF issued in 2015 an unambiguously negative opinion regarding its use by mBank, Alior Bank and Idea Bank, and the banks withdrew from this, despite the fact that this activity was not prohibited by the law at that time. Meanwhile, the United Kingdom, and then Lithuania, launched a so-called regulatory sandbox (a separated, safe environment in which it is possible to experiment without any legal or market consequences). As a result, after the announcement of Brexit, most online banks and payment services register in Lithuania, from where they operate across the whole of Europe on the basis of a European passport.

“An important barrier to development is also the fact that in the financial sector the view still lingers that people need personal contact with an advisor or salesperson in order to buy a new product or make an investment decision. Besides this, so far many institutions have treated fintechs as competition, rejecting the need to establish cooperation with external entities. Whereas companies with foreign capital, if they have implemented innovative solutions at all, had to use the products of their foreign head offices, so they could not cooperate with local suppliers,” adds Prof. Hryckiewicz.

However, the approach of both Polish and European enterprises is slowly changing. As analysts from both Deutsche Bank and Accenture stress, their management realize that failure to follow a trend means that sooner or later they will be excluded from the global market. According to the data of IDG, over three quarters of the forecast expenditure on AI and cognitive systems originates from American banks and trade. Today it is precisely the USA and China that have the most firms working on implementing artificial intelligence. Europe is to compete with them thanks, among others, to the EU plans to pump EUR20bn into innovative solutions in AI under the Horizon program.

In Poland, there is a large IT base for this. For example, in terms of contactless payments, pay-by-link payments, and sectoral solutions such as BLIK, Poland is among Europe’s technological leaders. Poland stands out in terms of express transfers, processes for remote sales of financial products (in the “pure online” formula), personal finance management (PFM), and remote customer service, e.g. in the form of video-consultation. Many of these solutions were introduced by banks in close cooperation with entities from the fintech sector.

One of the most important ideas is the announced launch by Alior Bank of one of the most modern innovation laboratories (iLab) in Europe aimed at improving customer experience, operating in the model of cooperation with customers and an accelerator for fintechs, based in the Open Api solution. It has already announced recruitment of start-ups from the fintech sector to the acceleration program. Alior Bank has also launched the platform Bancovo — the first transactional ecommerce in Poland offering cash loans of banks and loan companies. For financial institutions, using the potential of small innovators who were supposed to threaten them is the most intelligent strategy in order to beat them.

(, CC BY)

Otwarta licencja