In a pioneering study conducted by experts from Berkeley Haas and Harvard Business School, the influence of Generative AI on entrepreneurial performance among 640 Kenyan entrepreneurs was thoroughly evaluated.

The study, titled “The Uneven Impact of Generative AI on Entrepreneurial Performance,” explores the nuanced effects of AI-generated advice on small business revenues and profits, providing valuable insights into leveraging AI to support entrepreneurs, particularly in developing economies with limited access to personalized business mentorship.

Experiment Design and Context

The experiment involved two groups of entrepreneurs: a control group that received a standard business guide and a treatment group that had access to a GPT-4-powered AI business mentor via WhatsApp. This setup aimed to measure the tangible impact of AI advice on business performance metrics such as revenues and profits. WhatsApp was chosen strategically due to its widespread use in Kenya, making it an ideal medium for delivering AI-driven advice to a large and diverse group of entrepreneurs.

Key Findings

The study revealed a complex picture. On average, there was no significant overall impact of AI on business performance across the entire sample. However, this average effect masked substantial variations based on the entrepreneurs’ baseline performance levels:

– High Performers: Entrepreneurs who were already performing well before the intervention benefited significantly from AI advice, with reported improvements in business performance by over 15%. This subgroup effectively leveraged AI recommendations to enhance their business strategies and operations.

– Low Performers: Conversely, entrepreneurs who were struggling before the intervention saw their performance decline by approximately 8% after receiving AI advice. This suggests that the AI recommendations may not have been appropriately tailored to address the complex challenges faced by these entrepreneurs.

Both high and low performers actively communicated with the AI mentor, demonstrating a strong desire and willingness to seek advice. However, the types of questions asked and the nature of the challenges faced differed significantly between the two groups. Low performers tended to request assistance with more difficult and complex activities, which may have been beyond the AI mentor’s ability to handle properly. This imbalance between the sort of guidance required and the counsel offered is likely to have led to the reported unfavourable consequences for underperforming entrepreneurs.

The study underscores the potential of AI to support entrepreneurial growth while highlighting the importance of tailoring AI interventions to meet the specific needs of different entrepreneur groups to maximize positive outcomes.

Read More : Click Here