How to read these maps
Each map follows the same pattern:
- Regional context: a few dated, sourced facts about what is already happening in the Gulf.
- Use cases: what AI is commonly used for in this sector. Now means widely available, well-understood tools. Next means maturing, with careful pilots sensible. Later means emerging, worth watching rather than buying. These timings are our editorial judgement, not forecasts, and will vary by business.
- Start-ups and established firms: where each tends to have an advantage.
- Risks to manage and first steps.
A reminder: none of this is investment advice, and no business needs to do everything at once. The best opportunity is usually the problem you already understand best.
1. Real estate and property
Regional context
- Dubai Land Department (DLD) launched a pilot for tokenised real-estate ownership through the Prypco Mint platform in May 2025, open to UAE ID holders, with investments from AED 2,000 and transactions in dirhams. DLD projects that tokenised property could reach 7% of Dubai's market (about AED 60bn) by 2033. That is DLD's projection (DLD) [F].
Use cases
| Now | Next | Later |
|---|---|---|
| Bilingual listing descriptions and enquiry handling; document checklists for leasing; maintenance-request triage | Valuation support using transaction data (with a human valuer signing off); tenant-screening assistance within data-protection rules; energy monitoring of buildings | Agent-assisted transactions on tokenised platforms, within whatever rules regulators set |
Start-ups vs established firms: start-ups can build specialised tools (Arabic and English listing quality, maintenance platforms). Brokers and developers hold the relationships, the transaction history and the trust.
Risks: fair treatment of tenants and buyers (automated screening can carry bias); personal-data rules; overstated automated valuations; regulatory limits on tokenised products.
First steps: digitise leases and maintenance records; answer common enquiries with a supervised assistant; keep a person on every pricing and contract decision.
2. Logistics, ports and trade
Regional context
- Shipping through the Strait of Hormuz has been heavily restricted since the 2026 war began. An interim US–Iran agreement in June to reopen it broke down within weeks [fact base A1]. In response, Maersk has re-routed Gulf cargo through landbridges via Khor Fakkan, Fujairah, Salalah and Sohar, with Saudi cargo via Jeddah. It applies emergency freight surcharges and notes that insurers are reducing cover (Maersk, Operational Update 49, 23 Sep 2026) [F].
- The Red Sea was also effectively closed to commercial traffic in early 2026, and about 17 submarine cables pass through it [fact base A2].
Use cases
| Now | Next | Later |
|---|---|---|
| Shipment tracking and customer updates in Arabic and English; document extraction (bills of lading, customs forms); demand forecasting from sales history | Route and mode re-planning support when corridors change; agents that assemble customs paperwork for human approval; warehouse slotting optimisation | Multi-party agent coordination across shippers, forwarders and customs, subject to trust frameworks and regulation |
Start-ups vs established firms: start-ups can move quickly on visibility and document tools. Established forwarders and port operators hold the networks, licences and data.
Risks: decisions based on stale or incomplete data during disruption; dependence on a single cloud region; cargo and personal data crossing borders.
First steps: get shipment data into one place; set up alerts for route changes; plan contingency routes with people in the loop. Resilience comes before optimisation.
3. Banking, insurance and financial services
Regional context
- The Central Bank of the UAE's Guidance Note on AI (February 2026) sets expectations for licensed financial institutions. These include board accountability, an inventory of AI systems, bias testing at least annually, disclosures in Arabic and English, human oversight "in", "on" or "out of" the loop matched to risk, and the ability to stop using a system (CBUAE Rulebook) [F].
- In the DIFC, Data Protection Regulation 10 covers personal data processed by autonomous and semi-autonomous systems (Mayer Brown, Jan 2026) [F].
Use cases
| Now | Next | Later |
|---|---|---|
| Customer-service assistants with clear hand-off to people; fraud-alert triage; document processing for onboarding (KYC) | Credit and claims decision support with documented human oversight; compliance monitoring; personalised financial education | Supervised agents that execute routine servicing end-to-end within tight limits |
Start-ups vs established firms: start-ups (fintech, regtech) can serve compliance, e-invoicing and SME finance. Banks and insurers hold licences, customers and data, and carry the accountability.
Risks: unfair outcomes in lending or claims; explainability; concentration on a single AI provider (the CBUAE note addresses this); consumer-protection breaches.
First steps: build an AI inventory; classify each use by risk; test for bias; make sure every customer can reach a person.
4. Energy and utilities
Regional context
- AIQ announced a US$340m, three-year contract to deploy its ENERGYai agentic AI across ADNOC operations, covering more than 28 producing fields. ADNOC describes its aim as becoming the "most AI-enabled energy company" (ADNOC's own description) (AIQ, 10 Mar 2025) [F].
- The UAE's Barakah nuclear plant reached full four-unit commercial operation in September 2024, supplying up to 25% of the UAE's electricity needs according to ENEC [fact base K3.2].
- Globally, the IEA expects data-centre electricity use to roughly double by 2030 [fact base G1, G2]. That creates demand for power and cooling expertise in the Gulf.
Use cases
| Now | Next | Later |
|---|---|---|
| Predictive maintenance on pumps, turbines and grids; energy-use dashboards for buildings; safety document search | Agentic monitoring that flags and drafts responses for engineers to approve; grid balancing support as solar grows | Highly automated operations in specific, well-bounded facilities under strict safety cases |
Start-ups vs established firms: national energy companies drive the large programmes. SMEs and start-ups can supply niche sensors, inspection (drones), building-energy services and data-centre efficiency.
Risks: safety-critical decisions; cyber-security of operational technology; physical risk to facilities (demonstrated in 2026, fact base A3).
First steps: instrument what matters; start with maintenance prediction; keep engineers accountable for every safety decision.
5. Government and public services
Regional context
- UAE federal framework: 50% of government sectors, services and operations to move to agentic AI within two years, with phased rollout, assessment and training for all federal employees (Dubai Media Office, 23 Apr 2026) [F]. A federal Artificial Intelligence and Data Authority was approved in June 2026 (Dubai Media Office) [F].
- Abu Dhabi's Digital Strategy 2025–2027 (AED 13bn) aims at an "AI-native government" by 2027 [fact base UAE4]. Bahrain has a national AI policy for government entities [fact base BH1]. Oman has created an AI Special Zone in Muscat [fact base OM2].
Use cases
| Now | Next | Later |
|---|---|---|
| Bilingual service assistants; application pre-checks; internal document search | Agentic handling of routine applications with human decision-makers for exceptions; regulatory mapping (the UAE has approved a "Regulatory Intelligence" ecosystem, fact base K1.8) | Proactive services that anticipate citizens' needs, with consent and transparency |
Start-ups vs established firms: large integrators win the platforms. Specialist SMEs can supply Arabic language and accessibility services, testing and assurance, training and change management.
Risks: fairness and appeal rights; data protection; public trust; resilience of critical services.
First steps (for suppliers): understand procurement and data-residency requirements; show your governance and testing; invest in Arabic quality.
6. Retail, hospitality and consumer services
Regional context
- Measured generative-AI use in the UAE was 73.3% of the working-age population in June 2026, the highest in Microsoft's report [fact base UAE6].
- E-invoicing is mandatory in Saudi Arabia (ZATCA) [F] and is being phased in for UAE B2B transactions from 2027. B2C transactions are excluded from the current UAE scope (KPMG) [F].
Use cases
| Now | Next | Later |
|---|---|---|
| Bilingual customer chat and WhatsApp-style ordering support; product descriptions; stock forecasting; staff scheduling | Personalised offers within consent rules; supplier-ordering agents with approval limits; multilingual voice service | Shopping agents acting for customers, which will change how brands are found and compared |
Start-ups vs established firms: start-ups can offer plug-in tools for SMEs. Established retailers and family groups hold customer loyalty, locations and supplier relationships.
Risks: personal-data misuse; poor-quality Arabic; over-automation that loses the personal touch customers value.
First steps: clean your product and customer data; get e-invoicing right; try one supervised assistant for common questions.
7. Health and care
Regional context
- Saudi Arabia's Seha Virtual Hospital supports more than 242 hospitals across 48 specialties and 68 subspecialties, according to the Ministry of Health, which also reports its Guinness World Records recognition (MoH) [F].
- The Emirati Genome Programme reports close to 800,000 samples [fact base K1.3].
- In Abu Dhabi, a clinical trial of a brain–computer interface (Neuralink's UAE-PRIME, with Cleveland Clinic Abu Dhabi) is recruiting (ClinicalTrials.gov NCT06992596) [F].
- UAE law restricts health data from being stored or processed outside the country except in cases allowed by regulation (Federal Law No. 2 of 2019, Art. 13, per DLA Piper) [F].
Use cases
| Now | Next | Later |
|---|---|---|
| Appointment and referral handling; clinical note drafting for clinician review; bilingual patient information | Imaging and triage decision support under clinical governance; remote monitoring for chronic conditions | Genomics-informed personalised care at population scale |
Start-ups vs established firms: start-ups bring specialised clinical tools. Hospital groups and ministries hold patients, data and accountability, and set the standards.
Risks: patient safety; health-data residency; consent; over-reliance on unvalidated tools.
First steps: start with administration (scheduling, referrals); keep clinicians in charge of clinical decisions; confirm where data is stored.
8. Education and training
Regional context
- AI became a subject in UAE public schools from kindergarten to Grade 12 in 2025–26 [fact base UAE5].
- Saudi Arabia introduced an AI curriculum for more than six million public-school students in 2025–26. Its SAMAI initiative aims to train one million Saudis, and more than half a million had registered by August 2025 (Saudi Gazette) [F].
- The UAE framework for agentic government includes training for all federal employees (Dubai Media Office) [F].
Use cases
| Now | Next | Later |
|---|---|---|
| Lesson preparation support for teachers; bilingual practice tutors; corporate training content | Adaptive learning paths; skills assessment for workforce programmes; AI literacy for managers | Personal lifelong-learning companions tied to recognised credentials |
Start-ups vs established firms: edtech start-ups can supply Arabic-first content and tutoring. Schools, universities and training institutes hold accreditation and trust.
Risks: children's data; academic integrity; cultural and linguistic fit (research has found Western bias in Arabic-language model outputs, fact base K3.7).
First steps: teach staff first; choose tools with strong Arabic support and clear data terms; put learning outcomes ahead of technology.
Across all sectors: the quiet opportunities
Some of the most dependable openings are not sector-specific:
- Arabic-language quality: translation review, dialect support, accessibility.
- Data preparation: cleaning, labelling, migrating from paper.
- Assurance: testing, bias audits, documentation, AI inventories. Regulators such as the CBUAE now expect these.
- Training and change management: helping people adapt with confidence.
- Resilience: backups in more than one region, continuity planning and cyber-security.
Related
- Business overview · AI readiness checklist · Digitisation roadmap
- Illustrative archetypes (fictional examples of these ideas in practice)