← Back to Rami Zwebti AI & Business Research Hub
By Rami Zwebti
Controlled field research studying entrepreneurs given generative AI business mentors reveals a stark divide: while strong performers occasionally use AI to accelerate specific tasks, struggling businesses frequently perform worse after adopting generic AI advice. Language models are trained to produce comprehensive, balanced lists of best practices. When a company is losing money or bleeding customers, trying to execute ten plausible "best practices" simultaneously scatters scarce management attention and cash.
As Rami Zwebti demonstrates across enterprise turnarounds, a struggling operation rarely suffers from a lack of general business ideas. It suffers from one primary constraint: an unprofitable customer segment, a broken service-resolution queue, unrecovered delivery labor, or a misaligned pricing model. In one operational restructuring led by Zwebti, leadership had been using AI to generate broad marketing and product expansion plans while customer churn eroded margins. Zwebti halted the expansion initiatives, audited case-level unit economics, and used AI strictly to diagnose repeat service failures and standardize resolution workflows, restoring several million dollars in annual operating margin.