In Latin America, SMEs represent 99.5% of all businesses, yet 80% fail within their first five years due to poor cash flow management. Why is the banking sector still failing them? Discover how Banco de Crédito del Perú (BCP) shifted the paradigm by integrating Coinscrap Finance’s AI engine to turn raw transactional data into a powerful, predictive “Digital CFO.”
Welcome to the new era of personalized banking
In Latin America, micro, small and medium-sized enterprises (MSMEs) are not just another market segment; they are, quite literally, the driving force behind the region. According to official data from the Ministry of Production (PRODUCE), in Peru alone they account for 99.5% of the business sector and provide over 60% of formal employment.
However, behind these macroeconomic figures lies a silent and painful reality: 80% of these businesses fail before reaching their fifth anniversary. The main cause? Poor cash flow management and a lack of simple tools to understand their day-to-day financial health.
This is where the great paradox of the financial sector arises: whilst the owner of an SME spends sleepless nights balancing manual spreadsheets, their bank holds millions of pieces of transactional data that could save their business. The problem is that this data is trapped in a cryptic and unintelligible language.
Transactional “Noise”: The invisible barrier between the bank and SMEs
For a business owner, checking their corporate bank statement is often a frustrating experience. Bank transaction codes such as “TRF-V-99238-SUN” or “COMPRA-POS-LIM-22” offer no strategic value. They are simply data “noise”.
Traditional mobile banking tools for businesses have been designed as purely transactional channels (making transfers, paying salaries and checking balances). But the modern SME is no longer looking for a digital wallet; it needs a treasury advisor. Unable to find this at their trusted financial institution, business customers are forced to seek external accounting software solutions or, worse still, to operate blindly, increasing their risk of bankruptcy.
The key insight for banks: Offering credit is no longer enough to retain the business segment. The real value and competitive advantage today lie in offering management intelligence.
The BCP Case: Turning raw data into a ‘Digital CFO’
Aware of this structural gap in the Peruvian market, Banco de Crédito del Perú (BCP) —the undisputed leader with over 35% of the corporate banking market share— decided to change the game. Rather than developing a solution from scratch, which would have taken years, it partnered with Coinscrap Finance to integrate our COCO data intelligence engine into its corporate ecosystem.
The result? The launch of a state-of-the-art Business Finance Manager (BFM) that acts as a virtual Chief Financial Officer (CFO) for thousands of businesses.
By using Vertical Artificial Intelligence and Natural Language Processing (NLP) tailored to local conditions (automatically identifying taxes such as SUNAT, social security contributions and the Yape ecosystem), BCP’s BFM achieves:
- Cleaning and Enrichment: Convert the bank’s internal codes into clear business names and their official logos.
- Operational Categorisation: Automatically classify transactions in key business accounts (Suppliers, Payroll, Taxes, Logistics).
- Cash Flow Forecasting: Identifying recurring expenses to alert the business owner to future payment obligations before they affect their liquidity.
Results that speak the language of C-level executives
The results of this partnership demonstrate that looking after customers’ financial wellbeing is the best business strategy for the bank. The integration has enabled us to achieve metrics that have a direct impact on ROI:
- Unbeatable Accuracy: A categorisation rate of over 90%, specifically trained for the financial semantics of the Andean market.
- Multiplied Engagement: Business customers now access the app an average of 2.5 times more often, transforming mobile banking into a strategic daily reference tool.
- Intelligent Cross-selling: The bank now understands the ‘life stage’ of each SME individually, achieving a 15% increase in the conversion of financing and working capital credit products offered at precisely the right moment of need.
As Lorena Cépeda Aliaga, from BCP’s Product Management team, rightly points out: “Coinscrap has enabled us to provide our SME clients with a tool for managing their business finances, thereby helping us to connect with them and better understand their needs.”
The Future of Corporate Banking: From selling products to providing 24/7 advice
BCP’s success story serves as a beacon for the entire region. It demonstrates that the move towards Open Business Finance is not about launching complex features to meet regulatory requirements, but about simplifying life for business owners.
When a bank helps an SME understand its expenses and protect its cash flow, it ceases to be a generic service provider and becomes an indispensable partner. Data intelligence enables the bank to advise thousands of businesses in real time, 24/7 and at scale, without overburdening its dedicated team of executives.
The Spanish-speaking market has an urgent need for financial health, and the institutions that are first to transform their transactional data into treasury insights will be the ones to lead the next decade.
Would you like to find out more about the technical and methodological details of this project?
We have produced a comprehensive six-page technical report in which we detail the architecture, the project phases and the key lessons learnt from the roll-out of the BFM at Peru’s leading bank.
👉 Download the Whitepaper: BCP & Coinscrap Finance Success Case
Frequently Asked Questions (FAQs)
What is the real difference between a traditional PFM and a BFM for businesses?
Whilst a PFM (Personal Finance Manager) focuses on managing household budgets and personal savings (e.g. leisure, groceries, household bills), a BFM (Business Finance Manager) is designed to ensure the operational survival of a business.
The BFM analyses and categorises data according to a strict corporate taxonomy: national taxes and duties, payroll payments, payments to strategic suppliers and logistics costs (OPEX). Furthermore, a BFM does not merely aim to “save money”, but also to provide cash management tools and cash flow forecasting to prevent stock shortages or non-payment of debts.
How does the Coinscrap Finance engine achieve an accuracy rate of over 90% in local Latin American markets such as Peru?
General-purpose artificial intelligence systems often fail in local markets because they do not understand regulatory acronyms or local brand names. The Coinscrap Finance engine uses Vertical AI and Natural Language Processing (NLP) specifically trained for the Spanish-speaking financial sector.
In the case of BCP, the engine was calibrated to natively identify and interpret critical transactions such as tax payments to SUNAT, health contributions to EsSalud, and the bulk processing of micro-transfers within the Yape ecosystem, successfully filtering out the ‘noise’ from the transaction records and categorising them with 92% accuracy.
For a large-scale bank, how long does it take to integrate this technology, and what is the product’s time-to-market?
A recurring concern for Tier 1 and Tier 2 institutions is the slowness of in-house development, which typically takes between 12 and 18 months. Coinscrap Finance’s modular architecture, based on robust and secure APIs, allows data intelligence to be decoupled from the bank’s rigid core banking system.
By following the agile methodology applied to the BCP project (onboarding, local training and QA phases), Peru’s leading bank managed to launch its solution in record time—approximately 12 weeks—thereby reducing development costs and gaining an immediate competitive advantage.
How does the BFM affect the issuing bank’s profit and loss account (ROI)?
The return on investment is reflected in three measurable business areas:
- Operational efficiency: By enriching and cleaning up merchant descriptions (adding logos and real names), calls to the call centre regarding “unrecognised charges” are drastically reduced.
- Increased engagement: As demonstrated at BCP, the frequency of use of the business app increased 2.5-fold, boosting the value of the digital channel.
- Intelligent credit cross-selling: By mapping SMEs’ actual cash flow in real time, the bank can predict when a business will need financing (working capital) and launch hyper-personalised pre-approved offers, achieving increases of up to 15% in loan conversion rates.
