Between data releases and recommendation bans: Who drives investment decisions in Arab exchanges?

Arab capital markets have long monitored automation applications as auxiliary tools for examining financial data, but the landscape is shifting as algorithms move into investment recommendation seats and direct trade execution.Presence is no longer limited to automated screening; it has expanded to redraw the boundary between the human investor and automated systems.Through divergent regulatory pathways in Saudi Arabia, the UAE, and Egypt.

In the Saudi market, this shift is evident in both asset management and supervisory regulations; after the Capital Market Authority approved the “Sedco Capital Multi-Opportunity AI Fund” in February 2025, the market saw the launch of the “AI Algorithm Enhanced Saudi Trading Fund” in December 2025, whose net assets reached about 50.6 million Saudi riyals by the end of that year. The most notable development appeared on 24 September 2026, when the Authority released a draft of the regulatory instructions for algorithmic trading to solicit feedback, paving the way for implementation on 1 November 2026.

The Saudi project caps the proportion of algorithmic order messages to executed trades at 20 orders per executed trade during a trading day, excluding securities with very high liquidity. It also requires firms to impose strict controls, test systems before deployment, retain supervisory records and submit them to the Authority, and trade on demand. This coincided with the Authority’s August 2026 announcement that “Lomha Tannabou” had fulfilled the fintech-experience requirements to provide advice using artificial intelligence, alongside other firms listed in the fintech registry to offer automated consulting.

Conversely, Abu Dhabi focused on opening its data infrastructure to generative artificial intelligence; on 13 August 2026, the Abu Dhabi Securities Market became the first market in the MENA region to provide official real-time trade data through generative AI platforms, including ChatGPT and Claude. This linkage allows natural-language queries without requiring specialized professional trading platforms. It is matched by a regulatory framework from the Securities and Commodities Authority that defines an automated advisor as an algorithmic service that builds and rebalances portfolios according to goals and risks, obliging firms to disclose operational constraints and conflicts of interest, while the Abu Dhabi Global Market rules cover machine-learning models and large language models in digital investment management.

In Egypt, the General Authority for Financial Supervision set different boundaries in February 2026 that clearly separate digital access from decision-making intervention. While it allowed brokerage firms to market their services through certified digital platforms that meet encryption and data-protection standards, it prohibited those platforms from offering investment recommendations, ranking or favoring securities, or using forecasting or AI models to influence participant behavior, and also barred them from performing the core functions of a broker.

These divergences impose direct operational obligations on asset managers, brokers, and compliance officers across the region; in Riyadh, trading offices must review their algorithms to keep the twenty-order rate before the regulation takes effect and document software-test logs to avoid liability. In the UAE, portfolio managers face a reality where participants can query the market with direct textual questions, reducing data-access monopolies. In Cairo, financial institutions must limit software to improving user interfaces without slipping into prohibited predictive analytics.

This divergence aligns with the International OSCO Organization’s warnings about the risks of relying on third parties and complex data models;The real test goes beyond tool trials, placing institutions before the question of control and legal responsibility when automated estimates err.