286 AI use cases across nine business lines.

Most enterprise AI programmes are a list of pilots. This one is an operating portfolio — architected on a single ERP platform, governed centrally, and mapped to the P&L of every business line it touches.

Business context

Mwasalat Holdings is a UAE integrated-mobility group spanning bus networks, taxi fleets, passenger transport, and vehicle rental and leasing. Asset-heavy, public-facing, and operating in a market where service reliability is visible to regulators and riders alike. I joined as Group CIO in 2025, also advising the board on the Emirates Mobility acquisition.

The problem and the stakes

The group ran on fragmented systems across entities: separate ERPs, siloed fleet data, decision-making anchored in retrospective reporting. In mobility, that lag is money — vehicles dispatched against yesterday's demand, maintenance done on schedules rather than condition, pricing static while demand moves hourly. The board's mandate was not "do AI"; it was to make the group's decisions faster than its competitors'.

Decisions made

• One platform before intelligence. A One-Platform ERP consolidation (Oracle Fusion) first — because AI on fragmented data multiplies the fragmentation.
• Portfolio, not pilots. A portfolio of 286 use cases identified, assessed and prioritised across nine business lines — finance, supply chain, HR, service, sales and risk — each classified (generative, agentic, predictive) and tied to an owner and a measurable outcome.
• Physical operations as the anchor. Fleet IoT and telematics feeding predictive maintenance, demand-based dispatch, traffic-aware routing and AI dynamic pricing by demand and seasonality.
• Governance as a control tower. Human-oversight gates for high-risk use cases — dispatch decisions, biometric data — so engineering pace never outran public trust.

Where the portfolio landed

• Supply chain & manufacturing (114): predictive maintenance via IoT telematics, AI dispatch, right-sized parts inventory, dynamic routing and fleet balancing.
• Human capital (51): predictive attrition, AI skill-matching, intelligent shift planning for drivers and field staff.
• Service & CX (48): chatbots and intent engines, sentiment analysis, Level-1 call deflection, personalisation across B2C, B2B and B2G.
• Sales & leasing (47): dynamic pricing, upsell recommendations, accelerated B2B leasing and predictive vehicle availability.
• Finance, risk & EPM (26): predictive cash-flow forecasting, spend analytics with anomaly detection, automated invoice matching with fraud detection.

Outcomes

The shift the board can see: decisions moving from reporting to prediction. A single ERP backbone across entities. A governed AI portfolio in which every use case has an owner, a cost model and an oversight gate appropriate to its risk.

Lessons for other executives

First, sequencing beats ambition — the ERP consolidation was unglamorous and non-negotiable. Second, classification is governance: knowing which of your use cases are agentic tells you where your oversight gates go. Third, in asset-heavy businesses the highest-value AI is rarely a chatbot; it is the model that decides when a vehicle gets serviced and what a lease should cost today.

About these numbers

Role: Group CIO of Mwasalat Holdings (employer, 2025–present). The 286 figure is the count of use cases in the group's internal AI use-case register — identified, assessed and prioritised across nine business lines; a subset is in production, with the remainder staged on the roadmap. Category counts (generative, agentic, predictive) come from the same register, as at mid-2026. Financial and operational outcomes will be published as they mature and clear the group's confidentiality review.