Future Investment Initiative identifies expansion gap: only 23% of organizations move intelligent agents into operation

The figures released from Riyadh reveal a paradox governing today’s tech budgets: corporate spending on AI systems is set to reach $400 billion globally by 2026, according to IDC, and about 88 % of firms have deployed these tools in at least one function, up from 78 % in 2024. Yet the shift from pilot platforms to actual operational returns still faces a real bottleneck, as McKinsey research cited in the report shows that only 23 % of organizations have succeeded in scaling autonomous agents across a single work environment, even though 62 % are actively experimenting with them.

The institute launched its Future Investment Initiative report, published under the think tank arm as “Business Transformation: AI in Enterprises,” noting a shift in investment focus from fascination with models to redesigning daily workflow pathways.The equation is no longer tied to model experimentation, but to their ability to lower operating costs and generate directly measurable financial returns.Contributions from report partners, led by Brookfield, JD.com and Visa, show that the gap between isolated experiments and value creation requires redesigning data governance and human delegation levels.

In the realm of large-scale investment management, Brookfield identified over 800 distinct AI use cases slated for implementation in its portfolio and established a dedicated value-creation office comprising more than 30 specialists in investment, operations and technology. This perspective aligns with the extensive industrial deployment by Chinese e-commerce giant JD.com, which has embedded AI algorithms in more than 3,000 practical scenarios covering supply chains, retail, logistics, food delivery and healthcare.

The most pronounced shift toward autonomous agents appears in the payments sector; Visa operates hundreds of models in live production environments to support more than 100 financial products. The company's contribution hinges on a critical condition accompanying the agents' transition from merely answering queries to executing actual purchases: imposing strict spending caps, setting approval thresholds and defining precise authorizations that prevent unauthorized transactions.

This regulatory discipline reflects broader field concerns, as Deloitte research cited in the report shows that data privacy and cybersecurity are the top worries for roughly 73 % of participants, followed by legal, intellectual property and regulatory compliance risks at 50 %.Hardening infrastructure is no longer a regulatory luxury but the factor that determines operating cost and the resilience of technology systems against threats.

This scenario creates a direct imperative for executives and technology leaders in the Gulf and Egypt; the shift from consuming global models to deploying agents within corporate work environments calls for a reallocation of project budgets. The challenge is no longer purchasing language-model subscriptions, but building operational teams and internal financial governance that impose limits on autonomous software actions before they connect to sensitive databases or payment gateways. Those who today mandate clear delegation structures and strict return accounting for their systems are the ones who avoid the wasted-spending trap that past experiments fell into.