Fintech may be Saudi Arabia's first industry to deploy AI at real scale

Past the pilot stage It is worth being specific about what "at real scale" means here, rather than treating it as a marketing phrase: it means AI products with disclosed customer counts, disclosed transaction volumes, or funding rounds explicitly earmarked for scaling an AI product that is already generating revenue — not a roadmap slide describing what AI could eventually do for the sector. Most industries talk about AI in terms of pilots and roadmaps. Saudi fintech is one of the few local sectors where AI is already sitting inside live infrastructure that is processing real volume today, not a future promise. Three companies illustrate where that deployment is furthest along: credit underwriting, financial-crime detection, and the open banking rails that connect the two. Credit underwriting: Abwab.ai Underwriting is the natural starting point for this survey, since it is the layer with the clearest disclosed evidence of AI already running against real loan volume rather than a pilot. Abwab.ai raised a $4 million seed round — Speedinvest's first Saudi investment , with participation from MEVP — to build AI infrastructure specifically for SME lending. The company's positioning is notable because it is not a lender itself; it builds the underwriting layer that lenders plug into, and the underlying lending infrastructure it supports has already processed more than SAR 10 billion in loans. That is a meaningful amount of real credit decisioning running through AI-assisted infrastructure already, well before most industries have moved past experimentation. Financial crime and knowledge intelligence: Mozn On the compliance side, Mozn — a Saudi AI company that received a strategic investment from HUMAIN — has built out AI tools for financial-crime detection and knowledge intelligence, and reports more than 150 customers already using its products. Anti-money-laundering and fraud detection are naturally suited to AI: the underlying task is pattern detection across large transaction volumes, which is exactly what modern models are good at, and exactly the kind of workload banks and fintechs cannot practically staff with human reviewers alone as transaction volumes grow. The connective layer: open banking None of this works in isolation. Tarabut's open banking infrastructure — which the company says now handles more than 5 billion API calls — is the connective tissue that lets underwriting and fraud-detection tools actually see the transaction data they need, across multiple banks, in a standardized way. Tarabut's $50 million raise to deepen its Saudi push matters for the AI story specifically because AI underwriting and AI fraud detection are only as good as the data pipes feeding them, and open banking infrastructure is what makes those pipes possible at scale rather than through one-off bank integrations. Collections and customer operations: the next layer Underwriting and fraud detection are the most visible AI use cases in Saudi fintech today, but collections and customer operations are the logical next layer: once a lender has an AI-assisted view of who to lend to and can detect fraud on the transactions that follow, the same data and modeling approach extends naturally to predicting who is likely to miss a payment and automating the customer communication that follows. None of the companies covered here has disclosed products specifically in that layer yet, but the underlying infrastructure — underwriting models, transaction data pipes, fraud pattern detection — is largely the same infrastructure collections AI would need. Underwriting and fraud detection are, technically, the same problem It is worth being specific about why underwriting and fraud detection are the two use cases furthest along. Both are, at their core, pattern-recognition problems applied to transaction and behavioral data: underwriting asks "does this borrower's data resemble borrowers who repaid or borrowers who defaulted," and fraud detection asks "does this transaction resemble normal activity or an attack pattern." Machine learning models are well suited to exactly this kind of comparison at scale, which is why both use cases were among the first areas of finance anywhere in the world to see real AI deployment, and why two of Saudi fintech's most visible AI products — Abwab.ai's underwriting infrastructure and Mozn's financial-crime tools — sit in these two categories rather than in newer, less pattern-based areas of finance. The capital backing this shift is not incidental None of this AI deployment happens in a capital vacuum, and it is worth tracing exactly where the money funding it is coming from. The scale of capital flowing into Saudi SME lending specifically is part of why AI is showing up here first rather than in a slower-moving corner of fintech. TechScoop's look at why institutional capital is pouring into Saudi SME lending covers the broader shift of bank and institutional money into SME credit platforms like Lendo, and that