UAE insurer puts AI agents into production, cutting case time 50%
Al Wathba Insurance says AI agents now run case management in production, halving turnaround times.

· Last verified: 16 Nov 2026 (Core42)

Summary
- Al Wathba Insurance says AI agents now run case management in production, halving turnaround times.
- Quality checks jumped from a 20-30% manual sample to 100% of cases reviewed by agents.
- Models are served in-country by Core42 Compass, keeping data residency inside the UAE.
The latest
Case turnaround at Al Wathba Insurance has fallen by 50% after the Abu Dhabi insurer moved AI agents into live production across case management, the company announced. Cases are now read, classified, enriched and routed automatically, with agents assembling customer history a handler would otherwise compile manually. Chief Executive Shukri Salem Mussabih Almheiri said the aim is faster response and simpler insurance, not an added technology layer.
Details
- What is live: AI agents built on MagOneAI, the enterprise platform from UAE-headquartered Magure, handle reading, classification, enrichment and routing of customer cases and surface a next best action. Al Wathba says every consequential decision remains with a company professional rather than the system.
- The numbers: Case turnaround is down 50%, according to the company. Quality assurance coverage rose from a manual sample of 20-30% of cases to 100% reviewed, with agents also conducting and analysing customer satisfaction surveys in real time beneath the case management layer.
- Sovereign stack: The architecture operates within the UAE. The platform runs inside Al Wathba's own cloud environment, MagOneAI provides orchestration and governance, and Core42 Compass serves the models through private endpoints configured for in-country data residency.
- Regulatory frame: As a licensed insurer, Al Wathba is governed by the Central Bank of the UAE's Guidance Note on Consumer Protection and the Responsible Adoption and Use of AI and Machine Learning, which places accountability for AI outcomes at board level and requires fairness, explainability and human oversight.
- Controls claimed: Magure says human task nodes pause any workflow for review or override, with role-based access, monitoring and full audit trails across every use case. Personally identifiable information is anonymised before reaching any model, and prompts and outputs are not stored or reused.
- Vendor position: Magure Chief Executive Akhil Koka said regulated enterprises should not have to choose between moving fast and staying compliant, describing what is running as production-grade rather than a proof of concept. He said: "The hardest part is behind us."
- The economics: Chief Financial Officer Muralikrishnan R said the platform gives a scalable foundation reusable across multiple use cases instead of isolated solutions, improving the economics of future deployments. The company says current workflows were deployed without separate compliance workstreams for each.
- Core42's role: Core42, a G42 company specialising in sovereign enabled cloud and AI infrastructure, supplies the model serving layer. Vice President of Product Engineering Will Badr said separating model from orchestration lets the insurer swap models as the technology evolves without rebuilding applications or governance.
- The roadmap: Subject to governance, risk and deployment approvals, planned additions include a private AI assistant offering employees an approved alternative to public AI tools, followed by medical-underwriting support, first-notice-of-loss triage, and accounts receivable and payable automation. No timeline was given.
- The company: Al Wathba Insurance was established in 1996, is headquartered in Abu Dhabi, and is rated BBB with a Positive Outlook by S&P Global Ratings. Magure, registered in DIFC, says it serves more than 80 customers across banking, insurance, government and energy.
Background
Agentic AI refers to systems that execute multi-step workflows rather than answer single queries. For regulated insurers, the binding constraint has been supervisory expectations on explainability and human oversight, which is why most deployments have stopped at pilot stage rather than entering live operations.
Between the lines
The claim worth tracking is not the 50% figure but the jump in QA coverage from a 20-30% sample to every case. Sampling is how insurers have traditionally managed the cost of oversight. If coverage holds at 100%, it changes what regulators can reasonably expect of a licensed insurer. The reusable-foundation argument made by the chief financial officer points the same way: the cost of the next deployment, not the first, is the real test.
What's next
Watch whether the planned private AI assistant and medical-underwriting support clear governance, risk and deployment approvals, and whether Al Wathba discloses sustained turnaround and QA figures in future results or S&P rating reviews.
Source: Core42