AI-Driven Client Intelligence: Dubai Customs’ Customer Insight AI Platform
18 November 2026
By Juma Al Ghaith,, Chief AI Officer, Dubai CustomsEnabling trusted trade requires a detailed understanding of the businesses that sustain it. Their trading patterns, compliance records and growth trajectories inform how Customs designs services, manages risk and supports economic activity. Dubai Customs’ Customer Insight AI Platform (CIAIP), launched in 2026, brings this intelligence into daily decision-making. By combining AI-based segmentation with compliance assessment and a loyalty programme linked to trading performance, the platform supports the Department’s 2026–2030 strategy and its role as an economic enabler. It provides a basis for more responsive engagement with traders, while keeping regulatory compliance central to the allocation of programme benefits.
Built on Dubai Customs’ enterprise data infrastructure, the platform analyses more than 13.6 million customs declarations spanning five years, alongside records of fines, risk assessments, legal cases and active holds. It produces detailed profiles for more than 129,000 companies. Two complementary analytical pipelines apply Recency, Frequency and Monetary value (RFM) analysis to assess how recently and frequently companies trade, and the value associated with that activity. One pipeline measures customs revenue; the other uses the Cost, Insurance and Freight (CIF) value of traded goods. Both employ K-means clustering, an unsupervised machine learning technique that groups companies with similar characteristics into six segments, ranging from “Champions” to “Lost” customers.
Compliance scoring is integral to this analysis. Each company receives a composite score based on three weighted components: accumulated fines, risk assessment outcomes and legal case history. The average score across the active customer base is 96.7%. The platform also identifies licensing, regulatory, financial and operational holds. Assessing these factors alongside trading activity gives officers a more complete basis for engagement and programme assessment, ensuring that commercial value is considered together with adherence to customs requirements.
The platform classifies companies across six trade types: Import, Export, Transit, Temporary Admission, Transfer and Cargo Transfer. It also identifies potential participants in the Authorised Economic Operator (AEO) programme by cross-referencing companies in the Champions segment with excellent compliance scores and no existing AEO certification. As of May 2026, this analysis had identified 446 candidates for further assessment, providing a structured basis for expanding participation in the programme.
A More Responsive Approach to Client Segmentation
The platform applies machine learning to trade and compliance records to identify patterns across the customer base. Its six segments reflect similarities in company activity, giving officers a more detailed basis for engagement than transaction volumes alone. Daily processing updates the analytical view as trading patterns change, helping teams identify shifts in activity and determine where further engagement may be warranted. This supports a more proactive approach to customer service and a clearer understanding of the businesses trading through Dubai.
The segmentation pipeline is connected to Mirsal 2, Dubai Customs’ clearance system, and writes its results to a PostgreSQL database that supports the operational dashboard. Officers and account managers can filter companies by segment, compliance tier, trade type, AEO status, licence validity, hold status and account manager assignment. The platform also supports loyalty programme onboarding, announcements and bulk communications, allowing the Client Happiness Department to tailor its engagement to specific groups and their requirements.
From Intelligence to Incentive: The CAGR-Based Loyalty Programme
The Loyalty Points Programme translates customer intelligence into a financial incentive linked to trading performance and compliance. It provides for points to be redeemed against customs fees and participating government services, establishing a direct connection between the platform’s analysis and the benefits available to eligible traders.
Eligibility is assessed over a four-year period using the Compound Annual Growth Rate (CAGR) of each company’s CIF trade value. The system calculates CAGR between Years 1 and 3, projects the expected CIF value for Year 4, and compares that projection with the company’s actual performance. Where actual CIF exceeds the projection, the excess forms the basis for calculating loyalty points. This approach recognises performance relative to each company’s historical trading trajectory and gives the programme a consistent basis for assessing businesses of different scales.
The allocation incorporates two multipliers. A size multiplier accounts for differences in scale, with companies assigned to five tiers according to their Year 4 CIF value: Micro, Small, Medium, Large and Very Large. A compliance multiplier adjusts the allocation according to each company’s compliance score, linking the incentive directly to regulatory performance. The resulting adjusted excess CIF determines each eligible company’s share of a fixed annual pool, currently set at one million points.
For the 2025 analysis year, 15,651 of the 121,775 companies assessed qualified for the programme. Eligible companies recorded an average CAGR of 14.46%, with total adjusted CIF exceeding AED 494 billion. Points are published through the platform’s Loyalty Points module.
Technical Architecture and Evolution
The CIAIP uses a cloud-native, microservices-based architecture. Development has proceeded in stages, extending the platform from company segmentation and compliance analysis to loyalty administration, payment integration and targeted engagement.
| Phase | Date | Key Capability |
| 1.0 | Q3 2025 | Revenue-based company RFM segmentation using K-means clustering; compliance scoring based on fines, risk assessments and legal cases; classification across six trade types; daily processing using DataRobot, Oracle and Supabase. |
| 1.5 | Q4 2025 | Parallel CIF-based RFM segmentation; integration of licensing, regulatory, financial and operational holds; broker segmentation; identification of potential AEO candidates. |
| 2.0 | Q1 2026 | Launch of the CAGR-based Loyalty Points Programme; K-means scoring of Year 4 CIF values; size and compliance multipliers; allocation and publication of the loyalty pool; bilingual English and Arabic dashboard. |
| 2.5 | Q2 2026 | Integration of loyalty redemption with the customs payment gateway and public-sector partners; account manager assignment; communications tailored to customer segments; security hardening through DefectDojo and the continuous integration pipeline. |
Quantifiable Results
The platform’s analytical coverage and programme results demonstrate the breadth of information now available to Dubai Customs and its application across customer engagement, compliance monitoring and revenue analysis:
Customer Visibility: The platform has profiled 129,296 companies across five years of trading history, including 79,177 active within the 12-month monitoring window. Daily processing maintains an updated view of RFM segmentation, compliance scores and trade classifications.
Revenue Concentration: The Champions segment comprises 1,136 companies, approximately 0.9% of the profiled customer base, and accounts for 54% of customs revenue in the analysis. Loyal Customers, representing a further 4.2% of the base, contribute another 26%. These findings inform account management priorities and provide greater visibility into revenue concentration.
Compliance Monitoring: The average compliance score across the active base is 96.7%. Officers can identify companies with critical compliance scores and review licensing, regulatory, financial and operational holds within the same dashboard.
Loyalty Programme Reach: The 2025 assessment identified 15,651 eligible companies, with an average CAGR of 14.46% and total adjusted CIF exceeding AED 494 billion. The adjusted CIF figure is the model’s allocation measure, distinct from the monetary value of points awarded.
AEO Development: By May 2026, the platform had identified 446 potential AEO candidates through the combined assessment of segment, compliance performance and certification status, enabling more focused outreach and further evaluation.
Strengthening Customs–Trader Engagement
The Customer Insight AI Platform gives practical effect to Dubai Customs’ philosophy of enabling trusted trade. It connects an understanding of business activity with compliance assessment, allowing customer engagement and programme development to draw on a common analytical foundation. Its contribution to the strategy is tangible: more informed services for traders, closer attention to compliance, greater visibility into revenue patterns and a structured approach to identifying candidates for trusted-trader programmes.
The loyalty programme adds a further dimension by linking incentives to performance against a company’s projected trading trajectory. Incorporating compliance into the allocation reinforces the relationship between commercial development and responsible trading conduct. For Dubai Customs, this creates a basis for supporting business activity through Dubai while maintaining the standards that underpin confidence in its trading environment. Measuring the programme’s effect on participation, compliance and trading behaviour will be essential to assessing its longer-term contribution to the economy.
Dubai Customs is prepared to share the platform’s methodology, architecture and implementation experience with other customs administrations and international organisations. Its modular design separates data ingestion, analytical pipelines, scoring engines and the user interface, allowing individual components to be adapted to different data environments and customs management systems. The Department welcomes collaboration with the World Customs Organization (WCO) and member administrations seeking to develop their own client intelligence capabilities.
More Information
Juma.AlGhaith@dubaicustoms.ae