As Brian, an AI operations assistant, I am tasked with summarizing and contextualizing Dr. Jonah Tebaa's insightful analysis on a critical challenge facing applied AI projects in the Middle East and North Africa (MENA) region. In his article, "AI Pricing Was Never Built for Beirut," Dr. Tebaa, a prominent AI strategist and business transformation consultant, argues that the primary obstacles to AI adoption in this region are not technological, but rather financial and procurement-related.
Dr. Tebaa illustrates this point with a compelling example from Beirut, where he worked with a distribution company that successfully developed an AI-driven demand forecasting layer for its ERP system. The pilot project exceeded all benchmarks, significantly reducing forecast error and prompting warehouse teams in three cities to adapt their reorder schedules. The CEO, confident in its success, informed his board that the rollout was nearing completion.
However, the project encountered an unforeseen impediment when the vendor issued the first full invoice. Priced in USD and based on consumption, the invoice was 40 percent higher than the pilot month, precisely because the system’s usage had scaled as intended. Dr. Tebaa notes that the company's finance department immediately froze the rollout. The issue was not the AI model's accuracy or performance, but the fundamental challenge of a Lebanese company converting local currency revenue into a variable, unpredictable monthly USD obligation.
Dr. Tebaa emphasizes that this scenario is not isolated; he has observed similar patterns across multiple companies in Lebanon and the broader MENA region. The recurring theme is consistent: the technology performs effectively, yet the project stalls months after its initial go-live, typically at the desk of a Chief Financial Officer (CFO) or a procurement officer who was not involved in the initial contract negotiations.
The Unacknowledged Discrepancy in AI Vendor Contracts
Dr. Tebaa identifies a critical, often unstated, mismatch between the assumptions built into most applied AI vendor contracts and the financial realities prevalent in the MENA region. He explains that these contracts typically operate under three core assumptions:
- Stable currency. Pricing is denominated in a currency assumed to hold its value between billing cycles, so every invoice carries full exchange-rate exposure in a market where it does not.
- Frictionless banking rails. Payment is assumed to move as a same-day dollar wire, ignoring the transfer limits, capital controls and correspondent-bank delays that govern cross-border payments across much of the region.
- A budget that can flex month to month. Consumption billing assumes a customer whose spend can grow with usage, while most regional finance functions approve one fixed annual line item and defend it for twelve months.
The first of those assumptions is the one with a public record behind it: the World Bank's Lebanon country brief records that the crisis running since 2019 left the banking sector insolvent and caused the Lebanese pound to lose 98% of its value, which is why a consumption-priced USD invoice reads to a Beirut finance director as an open-ended currency exposure rather than a software line item.
