Saudi GovTech has moved from building portals to automating the work behind them

Count the Saudi government announcements at LEAP 2026 and the striking thing is how few of them are new portals. The portal problem was largely solved in the last cycle. What appeared this year addresses two different problems: too many portals, and too much manual work behind each one. Consolidation was the clearest theme. The Ministry of Media launched Ealam, a single portal folding in 22 separate media-sector platforms. That is not a new service. It is the removal of 21 front doors, and it is the kind of work that produces no launch excitement and a great deal of value, because every additional portal is another login, another support path and another integration that has to be maintained. Procurement structure was the second. SAMAI 2 sets an AI procurement framework across 11 ministries, which matters more than any single deployment underneath it. A framework determines what government entities can buy, from whom, on what terms and against which standards. Eleven ministries buying AI on a common framework is a different market for vendors than eleven ministries each running their own process, and it is the mechanism by which capability spreads rather than accumulating in whichever entity moved first. The third theme is the actual subject here: automating the processing rather than the presentation. Takamol's systems review more than 300,000 job-role requests a month using large language models and analytics, checking each against the licensing of the establishment making the request. SAL and Elm launched CargoGate to digitise shipment pickup and delivery at air-cargo stations. The Ministry of Justice unveiled AI-driven enhancements to its rights protection system. Salam and UnifyApps signed a memorandum on agentic automation for Saudi government and enterprise customers. Those are different in kind from a portal. A portal presents a service and collects an application. What happens next has traditionally been a person reading documents, checking them against rules and entering the result somewhere. That step is where the time goes, and it is the step the current wave of announcements is aimed at. The same pattern is visible at the physical border. The passports directorate demonstrated a palm-based contactless e-gate, and Tahakom showed Smart Pass biometric travel alongside its Siraj analytics. A biometric gate is workflow automation in a very literal sense: it removes the document check rather than digitising the form that precedes it. There is a standards layer forming under all this, and it is easy to miss. NAVER's digital twin platform was approved by the Digital Government Authority as a national standard for local government, connected to Balady. Approving a standard is a slower and more consequential act than launching a product, because every municipality that builds on it inherits the same data model. Standards are how the next five years of local government software gets shaped, and they are decided in announcements that read like vendor news. Now the discipline the category requires. The great majority of what was announced during the week is memoranda, framework agreements and pavilion demonstrations. The Ministry of National Guard signed multiple digital transformation memoranda. The Insurance Authority and HUMAIN signed a cooperation agreement to explore AI compute and secure data access. Civil Defence showed a smart helmet. Those are all genuine, and none of them is a service that a citizen or a business can use today. Two things announced around LEAP have numbers attached that survive scrutiny, and they are worth more than the rest combined. Absher, the interior ministry's platform, handled more than 240 million transactions across individuals and businesses in the first half of 2026. CargoGate launched after a pilot that recorded 92 percent operational adoption across SAL's main stations in Riyadh, Jeddah, Dammam and Madinah. One is a volume figure from a running system. The other is an adoption figure from a completed pilot. Both describe something that happened rather than something intended. The gap between those two categories is the honest summary of Saudi GovTech in 2026. The direction is right and the underlying platforms are unusually strong by international comparison, with national identity, payment and notification infrastructure that most governments do not have. What remains uneven is how much of the processing behind each service has actually been automated, as against how much has been announced as an intention to automate. There is a specific hazard in the next phase that the portal era did not have. Automating presentation is low risk, because a person still makes the decision. Automating processing means a system reads a document, extracts data and produces an outcome, and an error in that chain propagates without anyone seeing it. The systems being described do this at volume: 300,000 requests a month, hundreds of millions of transactions. Published accuracy rates, co