epeer

Epeer is developing a revolutionary behavioural AI-based credit scoring system that aims to assess the credit risk of young and unbanked people without a fixed income and extensive credit history to enable them to participate in the financial sector. We have built an innovative risk assessment tool with more than 90% accuracy of cluster prediction, determining the probability of future behaviour in terms of loan payment.

Epeer’s credit scoring technology is currently at a stage TRL 7. We developed and validated machine learning methods for data clustering/regression based on a metric called "repayment-ratio" defined based on user repayment information. Two deep network models were created with similar architecture and specifications but one for regression and clustering. The attributes were used to predict the coefficient of repayment rate. We have successfully conducted an internal pilot and proof of concept. We aim to train the algorithms on a bigger sample in order to validate and enhance the efficiency of the models.
Epeer is an actively functioning project in Poland and Ukraine, with a forthcoming entry on the Spanish and American market in the near future.

Our specific target clients are financial institutions which cannot properly assess credit risk of the young generation. Among these financial institutions, we can include banks, credit unions, credit scoring companies, governmental institutions, financial intermediaries. In the euro area (Eurozone), there are 4,954 monetary financial institutions, which belong to any of the following sectors: central banks, credit institutions, other deposit-taking corporations, money market funds. It is estimated that globally there are about 21,500 banks. Over 57,000 credit unions operate in 105 countries worldwide. The global market for credit risk assessing tools is still developing, yet there are no clear boundaries that allow for straightforward valuation. In order to estimate potential market size, we take global risks analytics valuation of $22,18 Billion in 2019 and use industry segmentation to focus on Banking, Financial Services and Insurance. This specific industry segment is the biggest one and responsible for circa 26% of total market valuation. Hence, the market size of credit risk assessing tools was around $5,77 Billion in 2019.
of risk analytics that is projected to reach $54,95 Billion by 2027.
There is limited data on the European market for credit risk assessing tools. Based on world lending data and the fact that the Western European Union is responsible for 35% of all loans given, we can assume that the total market value for European credit risk assessing tools was around $2 Billion in 2019.

Our product is targeted at young people, who do not have a credit history that would allow them to access financial services easily. Traditional financial institutions offer an outdated approach, which is not suited to assess the credit risk of those people in today’s environment. This client segment is growing exponentially and requires an innovative solution, created with up-to-date technologies that can overcome limitations of credit risk management. It is only possible with behavioural risk assessment and AI technology applied to today’s generations’ customs.
Young people in the EU are challenged with the problem of financial exclusion and liquidity. They are mainly caused by extensive consumption and emotional-driven spending under well-thought and well-targeted spending online and social media marketing strategies. This problem is extensively and constantly enhanced by unlimited and facilitated access to an enormous variety of goods and services worldwide, further aggravating the issue. When analysing the current market offerings to the younger generations (i.e., the target group with increasing demand for credit solutions) from banks and other financial institutions, there is a massive lack of available and, most importantly, suitable solutions. The younger generations remain the most neglected demographic groups due to the lack of credit history and fixed income. There is currently no available solution to overcome such deficiency, which seems to be the primary barrier to additional liquidity through traditional services. And primarily, young individuals have limited access to financial institutions. The scoring systems used by banks rely solely on traditional data, which causes millions of people in the EU to manage their finances through cash transactions with no physical tracks, even though they are active online users. There is a pressing need for alternative methods of credit scoring assessment, based on online behavioural data to serve untapped people and satisfy the market needs. Due to major regulatory issues, the European market is at an early stage of development in comparison to global standards and requires well-developed solutions, which can initiate the progress.

We are building an innovative AI-based credit scoring system for financial institutions, for the first time, enabling them to serve the young and the unbanked individuals. Epeer is developing a revolutionary behavioural AI-based credit scoring system that aims to assess the credit risk of young and unbanked people without a fixed income and extensive credit history to enable them to participate in the financial sector. We have built an innovative risk assessment tool with more than 90% accuracy of cluster prediction, determining the probability of future behaviour in terms of loan payment. The system analyses the cultural differences to identify the unified characteristics making the system self-adjusting to the changing environment. Our Artificial Intelligence system analyzes more than 1000 data points from traditional sources such as credit bureaus, open banking API. Still, more importantly, we focus on collecting behavioural data, such as social media activity, articles searched and read on the internet, hobbies and purchasing intentions, geolocation analysis, and more. This approach results in an extraordinarily low, 5% default rate, thus representing the most efficient risk assessment solution in the market. In summary, solving liquidity issues drives consumption and investments, contributing to higher GDP growth. Additional funds can stimulate higher education and make Europe more competitive worldwide with a bigger pool of skilful labour.

Currently, there are only a few companies that develop similar solutions on the European and global market. Those companies use different points as anchors for their credit risk assessing models. The selection of particular data used in training the algorithms significantly impacts the tool's accuracy. Epeer uses the weights' assessment to properly assess the correlation of the behavioural indicators with the credit repayment and potential future financial-related behaviour to mitigate the risk. A deep-tech AI-based scoring system with a self-learning neural network backed by a strong team of data scientists and AI experts makes Epeer stand out from the competition.

For the first time, using our revolutionary system allows us to determine individuals' future financial-related behaviour accurately. On top of that, the weights' assessment regarding the given data is also crucial in order to properly assess the correlation of the behavioural indicators with the credit repayment and potential future financial-related behaviour. On the other hand, the quality in terms of information accuracy of the sourced data is of great importance. Therefore a selection of reliable data providers with variable data determines the accuracy of the scoring model. In addition, the knowledge and expertise of AI experts and the appropriate selection of machine learning techniques allow for obtaining satisfactory results. A team of specialized data scientists is needed to process data efficiently. We are constantly looking for new sources of data to be included in our database that would positively impact the scoring accuracy. Based on the collected data, the system automatically assigns a potential borrower to a specific group of borrowers, corresponding to particular features and behaviours - thus clustering users. Our system exploits neural networks as classification tools and machine learning in the field of the track record validation and modification of classifiers. Our goal is to group people in such a way that our clients are assigned to a given cluster as close as possible to the estimated probability of repayment. It should be noted here that classifiers do not aim at searching within-person similarities but at classifying selected parameters and behaviours to isolate groups of people with similar financial behaviour. Our AI-based solutions allow us to transform credit scoring in many ways, making them more efficient and faster. With Epeer’s AI solution in credit scoring systems, financial institutions can get unique insights into their customers' financial behaviour based on collected data. Using our revolutionary system enables us to determine the probability of future behaviour in terms of loan repayment. Our scoring algorithm is not only innovative on the scale of loan companies. It also represents an undeniably competitive solid advantage of Epeer in the global financial sector.

We engage with financial institutions, primarily banks, by highlighting the untapped part of the market consisting of young people lacking fixed income. Among these financial institutions we can include: banks, credit unions, credit scoring companies, governmental institutions, financial intermediaries. Our revenue model is based on two pillars. The first one is licensing per inquiry. We provide customized risk analysis of a client to financial institutions. The second pillar is white labeling the lending platform for financial organizations to use at their own discretion, with their own branding for their own clients.

Epeer sets a differentiation strategy and offers a unique and highly specialized credit risk assessment tool, which allows us to offer premium services at a competitive price for the end customers. The value chain begins by sourcing data from external partners like credit information bureau or national banks. The second step is the suppliers of data which provide the behavioural data and open banking data. The inside operations are based on the AI-based analysis of the data, generating outputs that are then used to make decisions for credit institutions to assess the credit risk of individuals. The value chain can be defined as a new concept because Epeer introduces a new approach to credit risk assessment exploiting a unique set of data.

The funding from the programme will enable us to fully develop and commercialize our solution bringing a radical transformation to the financial market.
To commercialize our innovation, we need more specialists to analyze the enormous amount of data. In addition, we would like to understand the outputs of the neural network better. Therefore we need to employ an expert in neuroscience behaviour to adjust the grid. Likewise, it is necessary to engage more Machine/Deep learning specialists to manage the network layers. Another area of specialization is sourcing data. Therefore, the employment of skilled developers is crucial in the process and to modify and implement the front and back end of the platform internationally. In addition, as we are building the technology on the Microsoft Azure cloud, we need to employ specialists from this area as well. It is about the scale of the project.

Our significant risk of failure is associated with technology development. Our AI-based credit scoring systems and behavioural risk assessment technology require specialists and experts in the various high-technology fields. Moreover, our technology development is time-consuming. It demands ample cloud computing resources. There is also regulatory risk connected with changing the regulations and limiting access to behavioural and open-banking data. Another risk that may occur is a financial risk - the risk of insufficient funds to cover the specialists' compensation. Finally, further expansion on foreign markets requires multiple cross-cultural adjustments associated with consumer behaviour that need an extensive analysis in order to provide most effective results.

We didn't participate in any incubation/acceleration program.

Epeer has a long track of business and technology excellence and completed many milestones along the way. In 2019 Epeer’s was claimed as one of the twenty most breakthrough technological solutions globally in the financial sector by the most significant financial accelerator in Dubai. Microsoft chose Epeer as one of few technological startups for its prestigious Microsoft program. On top of that, Epeer was able to run a successful internal pilot and conduct a proof of concept for its technology. In 2021 Epeer finished selecting and validating machine learning methods and achieved 95% of cluster fit accuracy.

In the light of Polish (Patent Property Law) and European (European Patent Convention) regulations, it is not possible to patent an application (software) as an independent creation. The regulations in this respect are unambiguous and leave no room for interpretation.
On the basis of copyright law, however, all rights belong to Epeer sp. z o. o. The operating system, platform and algorithms developed until the beginning and during the whole process of development belongs entirely to the company. All the employees involved in the project are prohibited from competition in the form of confidentiality and non-compete clauses. The company has introduced rules and regulations which eliminate the risk of unwanted use or outflow of know-how and intellectual property rights belonging to Epeer. No outsider who has not been granted special access to the Epeer system has the possibility to view, learn, make changes, copy or distribute the technology that is used by the company. We conduct ongoing monitoring of competitive platforms, and our legal team will take appropriate legal steps in the event of any unauthorized use of epeer’s solution.