Artificial Intelligence is showing promise as a powerful financial forecasting tool for businesses in Ghana, especially small and medium-sized enterprises (SMEs) navigating today’s unpredictable economy.

A study led by Mr. Godwin Boakye Antwi, a chartered accountant and the current Budget Officer at Kwame Nkrumah University of Science and Technology (KNUST), Kumasi has found that an AI model known as Long Short-Term Memory (LSTM) significantly outperforms traditional methods commonly used by businesses.

Mr. Antwi’s research, which was published in the International Journal of Novel Research and Development (IJNRD), analysed three years’ worth of financial data from SMEs operating in retail, hospitality, and light manufacturing sectors within the Greater Accra Region.

He compared the forecasting accuracy of four different models, including the widely used ARIMA and other traditional tools, and found that LSTM produced the most reliable and precise results.

According to the study, LSTM was able to detect complex trends in revenue data that other models missed, offering more consistent and useful forecasts for business planning. This level of accuracy, the report noted, is critical for SMEs, which often face shifting customer behaviours, fluctuating costs, and tight margins.

One retail business owner quoted in the study remarked, “We’ve heard about these tools, but they’re expensive and not easy to connect to our current system,” highlighting the practical challenges businesses face in adopting advanced technology.

Despite its potential, the study revealed that most SMEs in Ghana have yet to embrace AI-based forecasting tools. Many cited the cost of technology, lack of technical expertise, and difficulties in integrating AI with their existing systems as major hurdles.

 Larger firms, those employing over 100 people were found to be more open to using AI forecasting tools, especially cloud-based platforms, although even they expressed caution about fully transitioning to these methods.

The research also pointed out how forecasting needs vary by industry. Retailers, for instance, use forecasts mainly for inventory and demand planning, while manufacturers focus on cost and supply chain management. Service-oriented businesses found forecasting more challenging due to irregular customer patterns and limited historical data.

Though the study focused on businesses in the Greater Accra Region, Mr. Antwi emphasized the need for further research in other parts of the country to understand the broader applicability of AI forecasting tools.