Tajikistan’s Zypl pioneers AI financial sector scoring system to drive SME lending

Tajikistan’s Zypl pioneers AI financial sector scoring system to drive SME lending
Tajikistan’s Zypl start-up is pioneering AI financial sector scoring system to drive SME lending and has already tied up with the country's third-largest bank and is now expanding / bne IntelliNews
By Ben Aris in Berlin June 6, 2023

Zypl, a software-as-a-service (SaaS) company founded in Tajikistan in 2021, is leveraging artificial intelligence (AI) models to drive financial inclusion in emerging and frontier market (EM/FM) economies.

Central Asia remains badly underbanked and the sector is still top heavy with state-owned banks focusing their attention on the big state-owned enterprises (SOEs). However, the bulk of economic activity remains concentrated in the small and medium-sized enterprises (SMEs) that employ the majority of the population. Zypl’s solution allows financial institutions, microfinance institutions (MFIs), and other financial enterprises to instantly underwrite loans for consumers with limited or no credit history.

Recognising that true financial inclusion goes beyond the quantity of loans provided to the underbanked. Zypl has developed proprietary technology that incorporates traditional social markers, behavioural analytics, web tracing and macroeconomic indicators to assess credit risk profiles.

The company already has a portfolio of $15mn of microcredits disbursed via 15 partners in five markets in and around Central Asia. The non-performing loan (NPL) rate on its scored loans is under 0.7%

Zypl's origins trace back to TajRupt, a non-profit organisation established by the same team a few years prior, reports Sturgeon Capital, a specialist emerging and frontier market fund. TajRupt focused on providing educational resources and extracurricular activities to Tajikistan students, helping them gain admission to prestigious US universities. In 2019, TajRupt received a grant from the Islamic Development Bank (IDB) to launch an AI Lab Academy, where high school students were trained in machine learning and AI optimisation frameworks and tools. However, upon graduation, these talented students faced limited job opportunities due to the lack of understanding and application of AI tools among enterprises in the country.

To showcase the potential of AI models in a business context, the team collaborated with two top MFIs in Tajikistan to develop a credit scoring algorithm for microloans. The project was successful, and the MFIs embraced the new credit scoring models. Subsequently Zypl was approached by Spitamen Bank, Tajikistan's third-largest corporate bank, to develop a similar product. Recognising the broader applicability of their solution, the founders decided to formally launch Zypl.

Since implementing Zypl's credit scoring solution, Spitamen Bank has expanded its product line to include microfinance offerings. It has issued $12mn in microloans, serving over 15,000 consumers who were previously underbanked or unbanked with no formal credit history.

Zypl currently collaborates with nearly two dozen partners in Tajikistan, Uzbekistan and Kazakhstan. The company is also expanding its coverage to financial institutions in Azerbaijan, Jordan and Kyrgyzstan.

Zypl continually enhances its algorithms, conducting regular bias testing to ensure fair decision-making and correlations. Additionally, the company works closely with regulators in EM/FM economies' central banks to influence policy changes that would further enhance credit scoring mechanisms.

Zypl's team strongly advocates for Open Banking and believes that establishing standardised API mechanisms for data sharing across utility, mobile and financial companies is crucial for refining credit scoring models. This approach would not only improve credit accessibility over time but also increase financial inclusion among previously unbanked and underbanked populations in EM/FM economies.