Case Study · Machine Learning
Asset Management & IFRS9 Loss Provisioning
JUMO turns the volatility and complexity of the African market into opportunity, but disbursing millions of dollars across multiple African markets quickly becomes risky, complicated, and nearly impossible to track by hand.
The Challenge
Volatile Markets. Millions on the Line.
JUMO operates at the intersection of scale and unpredictability, disbursing capital across multiple African markets where conditions shift quickly.
Forecasting cashflow and provisioning for credit loss in that environment is normally a people-heavy exercise. Often a five-person team dedicated to nothing else.
JUMO needed forecasting and provisioning models that could hold up against real market volatility, without needing to keep growing the headcount behind them every time a new market came online.
- Multi-market disbursement
- African market volatility
- Manual credit-loss provisioning
- IFRS9 compliance
- Cashflow predictability
Challenge
Solutions Built
Technology, Not Headcount.
Symbyte built two machine learning models: one forecasting JUMO’s financials six months out, the other automating credit-loss provisioning entirely.
Accounting Forecasts
Using Fourier orders, Symbyte predicts disbursements, collections, and profit and loss for the next six months within 95% accuracy, despite the volatility of the African market, giving JUMO predictable cashflow and a healthier business.
IFRS9 Expected Credit Loss Model
A custom Markov chain machine learning model predicts JUMO’s expected credit loss within 2% error, with no human intervention. Where other companies dedicate a five-person team to this calculation, Symbyte throws technology at the problem instead of people.
95% Accurate 6-Month Forecasts
Disbursements, collections, and P&L are forecast six months out within 95% accuracy, despite real market volatility.
Credit Loss Within 2% Error
The custom Markov chain model predicts expected credit loss automatically, with no human intervention required.
A Five-Person Team, Automated
What typically demands a dedicated provisioning team now runs as a model. Technology doing the work headcount used to.
Predictable Cashflow Across Markets
JUMO can now plan around dependable cashflow forecasts instead of reacting to volatility market by market.
