ZTABS builds production AI systems for Riyadh enterprises and government entities — Arabic-first NLP that handles Saudi dialect, computer vision for NEOM and smart city programmes, and enterprise AI that aligns with SDAIA governance frameworks and Vision 2030 localisation mandates.
Riyadh is in the middle of the most ambitious AI investment programme of any city on earth — and that is not hyperbole. Saudi Arabia's national AI strategy, overseen by the Saudi Data and AI Authority (SDAIA), targets AI contributing SAR 500 billion ($133 billion) to GDP by 2030. The Public Investment Fund (PIF) has allocated tens of billions to AI-related investments. NEOM — the $500 billion megaproject in Tabuk Province — is designed as an AI-native city where autonomous systems manage energy, transport, and logistics from day one. Riyadh itself is transforming: the Royal Commission for Riyadh City is executing a programme to grow the capital from 7.5 million to 15-20 million residents by 2030, with AI-powered urban management underpinning the infrastructure expansion. The AI development opportunities are enormous but the execution challenges are specific to Saudi Arabia. Arabic NLP remains significantly behind English and Chinese — Saudi dialect (Najdi Arabic in Riyadh, with significant Hejazi influence from the western province workforce) differs from MSA in vocabulary, grammar, and pronunciation. Most "Arabic AI" products are trained on MSA text from news corpora and fail on the conversational Saudi Arabic that actual customers and citizens use. Data localisation requirements under SDAIA's National Data Governance Interim Regulations mandate that certain categories of data be processed within Saudi Arabia — meaning cloud-first AI architectures that default to US or European data centres need redesign. Saudisation requirements (Nitaqat programme) mean AI projects must plan for Saudi national workforce development, not just deploy foreign technical teams. And the procurement environment — dominated by large government entities, semi-government organisations like ARAMCO, SABIC, and STC, and PIF portfolio companies — requires navigating NUPCO, Etimad, and other government procurement platforms that have their own technical and compliance requirements.
Riyadh's AI development demand is driven by government digital transformation, energy sector modernisation, financial services, and healthcare — each with billion-riyal budgets. Government and public sector: Saudi Arabia's government entities are the largest AI buyers in the region. SDAIA itself operates the National Centre for AI (NCAI) and commissions AI projects across government. The Ministry of Interior, Ministry of Health, Ministry of Education, and 20+ other ministries have AI transformation mandates. The Zakat, Tax and Customs Authority (ZATCA) deploys AI for tax compliance and customs processing. Absher (the government services platform serving 28 million users) uses AI for document processing and service routing. The General Authority for Statistics uses AI for census and economic data analysis. Government AI procurement flows through the National Centre for Government Resource Systems (NCGRS) and Etimad platform — understanding procurement requirements (Saudisation percentages, data localisation attestations, SDAIA governance alignment) is essential. Energy and mining: Saudi Aramco — the world's most valuable company — operates AI programmes spanning upstream exploration (seismic interpretation, drilling optimisation), midstream operations (pipeline integrity monitoring, refinery process optimisation), and downstream distribution. SABIC applies AI to chemical manufacturing process control. Ma'aden (Saudi Arabian Mining Company) uses AI for geological modelling and processing plant optimisation. The Ministry of Energy's Renewable Energy Project Development Office (REPDO) manages Saudi Arabia's solar and wind programme — AI for energy forecasting and grid integration is a growing requirement. Financial services: SAMA (Saudi Central Bank) regulates 30+ banks including Al Rajhi Bank (the world's largest Islamic bank by assets), Saudi National Bank, Riyad Bank, and Banque Saudi Fransi. The Saudi fintech sector has grown to 200+ licensed companies under SAMA's Fintech Saudi initiative. Tabby, Tamara, and STC Pay handle billions of riyals in digital payments. AI applications: Shariah-compliant credit scoring (cannot use interest-based risk models), fraud detection for the rapidly growing digital payment ecosystem, and regulatory reporting automation. Healthcare: the Ministry of Health operates 270+ hospitals and 2,300+ primary care centres. SEHA (the e-health system) digitises patient records across the kingdom. King Faisal Specialist Hospital & Research Centre, King Abdulaziz Medical City, and private hospital groups (Dr. Sulaiman Al Habib, Saudi German Hospitals) invest in diagnostic AI, clinical decision support, and operational optimisation. The National Health Information Centre (NHIC) manages health data governance — AI projects accessing patient data require NHIC coordination.
AI development in Riyadh requires mastering four dimensions that are unique to the Saudi market. Arabic NLP with Saudi dialect competence: standard Arabic NLP models (AraBERT, CAMeL tools) are trained primarily on MSA text. Saudi conversational Arabic — as used in customer service interactions, social media, and government citizen communications — contains vocabulary, grammatical constructions, and code-switching patterns these models handle poorly. We fine-tune Arabic language models on Saudi-specific corpora: customer service transcripts, social media data from Saudi Twitter (one of the world's highest per-capita Twitter usage rates), and government citizen interaction logs. Our models handle Najdi dialect features, Saudi-specific terminology (government programme names, local business terms), and Arabic-English code-switching common in Saudi professional communication. For document processing, we handle Arabic's right-to-left text direction, Arabic-numeral mixing (SAR amounts in Arabic text), and the various Arabic font rendering challenges that break standard OCR. SDAIA governance alignment: every AI system deployed in Saudi government or handling Saudi citizen data must align with SDAIA's governance frameworks. This means: data classification according to NDMO's National Data Governance Interim Regulations (public, restricted, confidential, top secret), algorithmic impact assessments for systems affecting citizen services, model documentation satisfying SDAIA's AI ethics principles (fairness, transparency, accountability, privacy, security), and data localisation compliance — ensuring Saudi data is processed on infrastructure physically located within the Kingdom (typically on Saudi Cloud Computing Company infrastructure, Oracle Cloud's Jeddah region, or AWS's upcoming Riyadh region). We embed these requirements into our ML pipeline templates so governance documentation is generated automatically alongside model training. Saudisation and knowledge transfer: government AI contracts require meaningful Saudisation — not just a Saudi project manager but actual transfer of AI engineering capability. We structure projects with explicit knowledge transfer workstreams: Saudi engineers embedded in the development team from week one, documentation in Arabic, and a handover plan that enables the client's Saudi team to operate and maintain the AI system independently. This is both a contractual requirement and good engineering practice — systems that the client cannot maintain in-house become expensive dependencies. Infrastructure and deployment: Saudi Arabia's cloud infrastructure is maturing but not yet at parity with US or European regions. We design for hybrid deployment — sensitive workloads on-premise or on Saudi Cloud infrastructure, with less sensitive processing on hyperscaler clouds where Saudi regions are available. For ARAMCO and defence-adjacent projects, air-gapped deployment with no external connectivity is standard.
Government / Public Sector
A Riyadh-based government entity processing 1.8 million citizen service applications annually (permits, certificates, complaints, information requests) relied on manual review by 280 processing officers. Applications arrived through multiple channels: the Absher-linked digital portal (62%), walk-in service centres (23%), and email/fax (15%). Documents were in Arabic (85%), English (10%), and mixed Arabic-English (5%). Processing time averaged 8.3 working days per application. The entity had attempted automation twice: a commercial RPA vendor's solution handled structured digital forms but could not process handwritten Arabic applications or scanned documents. An international AI vendor's document processing system achieved 71% accuracy on Arabic text — too low for government service delivery, where errors affect citizen rights. Both solutions failed on Saudi-specific document formats (national ID copies, commercial registrations from the Ministry of Commerce, Iqama documents for resident workers).
We built an end-to-end document intelligence system with five components: (1) Multi-channel intake normalisation: converting all application formats (digital forms, scanned paper, photographs of documents, email attachments) into a standardised processing format with consistent metadata. (2) Arabic-first OCR and document understanding: a custom vision-language model trained on 45,000 annotated Saudi government documents — national IDs, commercial registrations, Iqama cards, educational certificates, and the entity's specific application forms. The model handles handwritten Arabic (a particularly challenging OCR task due to connected script and diacritical marks), printed Arabic in multiple fonts, mixed Arabic-English documents, and poor-quality scans from walk-in centres. Accuracy: 94.2% on printed Arabic, 87.6% on handwritten Arabic — both above the 85% threshold required for automated processing with human verification of low-confidence extractions. (3) Application classification and routing: a text classification model that reads the application content and routes it to the correct department and service category — 47 service types across 12 departments. Classification accuracy: 96.1%, with confidence-based escalation to human officers for ambiguous cases. (4) Completeness checking: automated verification that all required documents are present and valid — checking that the national ID number matches across documents, that commercial registrations are current, and that required attachments are included. Missing document notifications are generated automatically in Arabic, reducing the back-and-forth that extended processing times. (5) Priority scoring: a rules-based system (not ML — SDAIA governance review determined that citizen service prioritisation should use transparent rules, not opaque models) that flags urgent cases, identifies VIP routing requirements, and ensures SLA compliance for time-sensitive applications.
Average processing time dropped from 8.3 working days to 2.1 working days — a 75% reduction. Applications requiring zero human intervention (fully automated end-to-end): 34% of volume, primarily standard digital submissions. Applications requiring minimal human verification (automated processing with officer confirmation): 48% of volume. Complex cases requiring full human review: 18% of volume (down from 100%). Processing staff redeployed from 280 to 180, with 60 officers retrained as AI system operators and quality auditors — satisfying the knowledge transfer and Saudisation requirements. 40 officers transferred to the expanded customer service function (handling the complex 18% of cases with more attention). Annual cost reduction: SAR 18.2 million in processing efficiency. Citizen satisfaction scores for processing speed increased from 3.1/5 to 4.4/5. The system processes approximately 7,200 applications per working day and has maintained accuracy above 93% across 14 months of operation.
Integrate GPT-4, Claude, Gemini, or open-source models into your product with custom fine-tuning for your domain.
Autonomous agents that handle complex workflows — lead qualification, support, data processing, and operations.
Turn your documents, data, and knowledge base into searchable AI-powered systems with accurate, cited answers.
Custom ML models for forecasting, classification, anomaly detection, and recommendation engines.
Image analysis, document understanding, text classification, sentiment analysis, and entity extraction.
Smart search, content generation, summarization, personalization, and intelligent recommendations for your SaaS.
Riyadh's AI development market is shaped by government procurement dynamics, international vendor presence, and a growing local ecosystem. (1) Global consultancies and system integrators: Accenture, IBM, McKinsey (QuantumBlack), Deloitte, and PwC have significant Riyadh operations — government digital transformation contracts worth billions of riyals flow through these firms. They provide programme management and strategy but typically subcontract technical AI development to specialist firms or their offshore delivery centres. Rates: SAR 2,500-6,000/hr for senior consultants. (2) International AI specialists: Palantir has Saudi government contracts for data analytics and security AI. C3.ai and DataRobot have entered the Saudi market for enterprise AI platforms. Google, Microsoft, and AWS offer AI services through their Saudi partner networks. These provide platform capabilities, not custom development for Saudi-specific requirements (Arabic NLP, Saudi document formats, SDAIA governance). (3) Regional technology firms: Elm (a PIF portfolio company) is the largest Saudi IT services provider, operating government platforms including Absher. Thiqah (another PIF subsidiary) handles government digital services. SITE (Saudi Information Technology Company) provides technology services to government entities. These firms have procurement relationships but varying AI depth. (4) Growing local AI ecosystem: Mozn (Riyadh-based, raised $10M+ in funding) builds Arabic AI products for financial services. Lucidya provides Arabic social media analytics. HAZEN.ai offers computer vision for smart city applications. The AI ecosystem in Riyadh is young but growing rapidly, supported by SDAIA's AI accelerator programmes and PIF investments. (5) Indian IT services firms: TCS, Infosys, Wipro, and HCL have expanded Saudi operations to capture government IT spending. They offer competitive rates (SAR 400-800/hr) but face Saudisation percentage challenges and often lack Arabic language AI expertise. ZTABS positions between global consultancy pricing (SAR 2,500-6,000/hr) and Indian IT services rates — delivering Arabic-competent AI development with genuine SDAIA governance expertise at SAR 800-1,400/hr, with Saudisation-compliant team structures and knowledge transfer built into every engagement.
AI development in Riyadh operates within government procurement frameworks for public sector work and competitive market rates for private sector. Government AI projects (typical ranges): AI proof of concept with SDAIA governance documentation (8-12 weeks): SAR 350,000-650,000 ($93,000-$173,000). Production AI system with Arabic NLP capability (16-24 weeks): SAR 700,000-1,800,000 ($187,000-$480,000). Enterprise AI platform with knowledge transfer programme (24-36 weeks): SAR 1,800,000-4,500,000 ($480,000-$1,200,000). Government projects typically include 15-20% budget allocation for Saudisation and knowledge transfer activities. Private sector AI projects: Proof of concept (6-10 weeks): SAR 250,000-500,000 ($67,000-$133,000). Production system (12-20 weeks): SAR 500,000-1,500,000 ($133,000-$400,000). Enterprise platform (20-30 weeks): SAR 1,500,000-3,500,000 ($400,000-$933,000). The government entity's SAR 18.2M annual savings against approximately SAR 2.8M development investment (6.5x first-year return) demonstrates the ROI for well-scoped government AI projects. For ARAMCO and large semi-government organisations, AI projects regularly exceed SAR 5M given the scale of operations and data volumes. ZTABS rates: SAR 800-1,200/hr for AI engineering, SAR 1,200-1,800/hr for ML architecture and strategy. Below global consultancy pricing while delivering implementation capability that strategy firms subcontract out.
Verified reviews from real client engagements — sourced from our public testimonial archive and Clutch profile.
My experience is throughout positive. Communication, service, the short response times and the flawless execution of a challenging topic was absolutely great. ZTABS is definitely my first choice again.
Christian Neff
Bank Software Advisory · Bank Software Advisory
Fintech
Fantastic Agency! I couldn't fault them even if I tried. They always go above and beyond to meet your expectations and always produces quality work. Thank you ZTABS.
Stephanie Kal
CEO · Beauty Finder Australia
Marketplace
It has been great working with ZTABS. They bounce off the ideas along the way. Amazing Experience.
Joel Rowe
CEO · Drill Quoter
Marketplace
We don't just contract. We ship and operate our own software. 23 products in production.
Common questions about AI development for Riyadh businesses
Yes — this is a core differentiator. Standard Arabic NLP models trained on news text perform poorly on Saudi conversational Arabic. We fine-tune on Saudi-specific corpora including customer service transcripts, Saudi social media, and government citizen interactions. Our models handle Najdi dialect features, Saudi terminology, and Arabic-English code-switching common in professional Saudi communication.
ZTABS is a remote-first AI development agency serving Houston businesses — including llm integration & fine-tuning, ai agents & automation, rag & knowledge systems. We work with Energy & Oil/Gas, Healthcare & Biotech, Aerospace & Defense companies in Houston, TX via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
AI Development in New York, NYZTABS is a remote-first AI development agency serving New York businesses — including llm integration & fine-tuning, ai agents & automation, rag & knowledge systems. We work with Finance & Fintech, Media & Advertising, Fashion & Retail companies in New York, NY via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
AI Development in Los Angeles, CAZTABS is a remote-first AI development agency serving Los Angeles businesses — including llm integration & fine-tuning, ai agents & automation, rag & knowledge systems. We work with Entertainment & Media, E-commerce & DTC Brands, Gaming & AR/VR companies in Los Angeles, CA via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
Web Development in Riyadh, Saudi ArabiaZTABS is a remote-first web development agency serving Riyadh businesses — including full-stack development, progressive web apps, api development. We work with GovTech, FinTech, Energy Tech companies in Riyadh, Saudi Arabia via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
Web Design in Riyadh, Saudi ArabiaZTABS is a remote-first web design agency serving Riyadh businesses — including ui/ux design, responsive design, custom interfaces. We work with GovTech, FinTech, Energy Tech companies in Riyadh, Saudi Arabia via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
Digital Marketing in Riyadh, Saudi ArabiaZTABS is a remote-first digital marketing agency serving Riyadh businesses — including search engine optimization, pay-per-click advertising, social media marketing. We work with GovTech, FinTech, Energy Tech companies in Riyadh, Saudi Arabia via timezone-aligned engineers and async workflows; we do not have a local office, and we are explicit about that with every client.
AI DevelopmentLearn more about our AI development services nationwide.
Partner with ZTABS for expert AI development in Riyadh. Get a free consultation today.
