Organisations in the UAE and Saudi Arabia are increasingly embracing artificial intelligence to transform customer and citizen engagement, but significant challenges remain in moving from isolated AI applications to large-scale, integrated deployment, says a new study.
The study by global cloud communications platform Infobip highlights the gap between growing adoption and organisational readiness, pointing to infrastructure limitations, data governance and the need for human oversight as key factors shaping the next phase of AI-driven engagement.
Drawing on the survey of 159 enterprises in the UAE and Saudi Arabia, the study found that 75% have implemented customer-facing AI agents, rising to 93% when organisations planning deployment within the next 12 months are included. Yet only 10% are scaling governed agentic AI across journeys with human oversight. Overall, 63% say their architecture limits AI to selected channels or otherwise constrains deployment at scale, while 51% require in-country or GCC data storage and processing.
Key findings at a glance:
Adoption is uneven across the maturity curve: 29% are still exploring or evaluating AI use cases, 15% are running pilots, 18% use AI in production within a single channel, 26% use AI in production across multiple channels, and only 10% are scaling governed agentic AI with human oversight.
About 62% have not yet deployed AI across multiple channels; 90% have not yet reached governed scale. 44% cite interaction volumes that human agents alone cannot handle as their leading driver for AI adoption, followed by rising customer expectations for faster, personalised service (38%).
As many as 32% cite a national AI strategy or government digital transformation mandate as a factor, reflecting the wider push toward digital government across the Gulf.
"The GCC has no shortage of AI ambition. The challenge is turning that ambition into trusted service at scale," said Zeid Shubailat, Regional Director at Infobip. "An AI agent is only as dependable as the operating environment around it. Businesses need one layer that can connect channels, preserve context, enforce guardrails and make every handover accountable. With those foundations in place, organisations can expand autonomy with greater confidence while continuing to meet sovereignty, security and service-quality expectations."
Building the operating layer for trusted AI
Infobip recommends that GCC organisations focus on three practical priorities as they move beyond isolated pilots.
* Unify context across channels. AI agents and human teams should work from a common customer identity, interaction history and set of policies. This allows journeys to continue across messaging, voice and human support without forcing customers to repeat information.
* Design governance and sovereignty into the architecture. Organisations should define which data AI can access, which actions it can take independently, when human approval is required, where data is stored and processed, and how every decision will be traced.
* Measure service quality and resilience. AI programmes should be assessed against customer satisfaction, resolution performance, handover quality, compliance and continuity, not efficiency alone.
Connecting AI, communications and human support
Infobip's AgentOS supports this operating model by bringing AI agents, journey orchestration, conversational customer data and human-led support together across one platform. Organisations can activate agents across messaging, voice, email and other channels, carry context between interactions, apply operational guardrails, escalate complex cases to people with the conversation history intact, and monitor performance across the journey.
The platform combines generative AI, conversational data intelligence and omnichannel messaging within a single orchestration layer. This helps organisations move from separate channel deployments towards connected experiences in which automation and human expertise can work together. It also gives business and technology teams flexibility to introduce AI according to their priorities, technical environments and governance requirements.
For organisations in the GCC, this approach can support a gradual path to scale: begin with clearly defined journeys, assign accountable owners, establish data and approval rules, and widen agent autonomy only when controls and service performance have been demonstrated, the company said.
"The objective is not autonomy for its own sake, but reliable digital service that preserves trust as interaction volumes and customer expectations grow," it added. -TradeArabia News Service