The gap between Saudi AI spending and returns comes down to governance and scale -Liferay

Saudi companies have moved past AI pilots, but the gap between what they spend on AI and what they get back hasn’t closed, and digital experience platform provider Liferay thinks it knows why. Regional Manager Ahmad Saad tells EnterpriseAM that “the biggest gap is between experimentation and seamless operationalization.”

Running a tool inside one team is easy. The harder part is scaling it across the enterprise while maintaining security, governance, and data quality. That, Saad said, is where the return on AI will increasingly be decided — less by how much a company spends on models than by how well it connects AI to the data and systems it already has.

Governance is both the company’s sales pitch and, Saad argues, the market’s blind spot. When departments adopt AI on their own, he said, organizations end up with “different tools, data sources, and security approaches,” duplicating cost and leaving IT and leadership unclear on where AI is running and what it can reach. Liferay’s answer is AI Hub, a low-code platform for building and running AI agents on top of the access controls of its digital experience platform. On it, agents act on behalf of authenticated users and reach only the data those users are cleared to see, with interactions logged. It’s model-agnostic and connects to outside systems through the Model Context Protocol (MCP).

SOUND SMART- Chatbots and enterprise agents are not the same. “A chatbot primarily responds to a user, while an enterprise AI agent participates in a business process,” Saad said. A support agent, in his example, should do more than answer a question — it could access authorized customer information, identify the issue, retrieve the relevant knowledge, and support the next step in the workflow, all within set permissions. At enterprise scale, organizations need to see what an agent accessed, what it did, and under whose authority it acted, he said.

In large organizations, the issue is less a lack of data than the inability to make the right data “accessible, connected, and governed,” Saad says. Companies hold information across websites, customer platforms, knowledge bases, enterprise applications, and systems of record. “If those systems remain disconnected, AI only sees part of the information required to complete a task effectively,” he tells us.

Data volume decides who pulls ahead: Saad identified sectors with high data volumes and heavy customer or employee interaction as best positioned to benefit, including financial services, telecoms, government, healthcare, retail, and travel and tourism. The biggest financial upside, he noted, likely lies in use cases where agents support several steps of an existing process — from customer-service triage to content pipelines, compliance review, and supply-chain risk monitoring.