Transforming Dubai Airport with intelligent contextual reasoning
Dubai Airport partnered with Nucleus to build an AI system that understands complex passenger inquiries, delivers personalized recommendations, and supports airport operations — going far beyond traditional chatbots.
Dubai Airport, a global leader in international travel, set out to enhance passenger experiences and optimize operations with AI. Nucleus partnered with Dubai Airport to develop a sophisticated system capable of understanding complex passenger inquiries, delivering personalized recommendations, and providing intelligent support across airport operations.
Beyond traditional chatbots and RAG
Traditional chatbots and retrieval-augmented generation (RAG) systems often fail to grasp the nuance and context of intricate passenger queries. They return generic, less relevant answers that don't meet a traveler's specific needs. Dubai Airport needed a system that could reason across flight times, locations, dietary needs, mobility restrictions, and live congestion — not just match keywords.
Recorded December 2024 — an early demo of Nucleus's contextual reasoning versus a standard chatbot, captured before the latest wave of reasoning models.
How Nucleus solved it
Multi-LLM ensemble
Rather than relying on a single model, Nucleus's brain-capacity system integrates multiple best-in-class LLMs, dynamically routing each query to the most suitable model based on domain and intent.
Multi-hop reasoning
When a passenger asks about dining, the system weighs their location, flight time, dietary needs, mobility, and current congestion to surface the most suitable recommendation — multi-step inference conventional bots can't do.
Human intelligence augmentation
The neural network is enriched with knowledge from domain experts, letting the AI handle a broader range of inquiries with greater nuance and empathy.
Contextual image integration
Computer-vision techniques fetch the most relevant images — maps, facility photos, product images — from Dubai Airport's repository to enrich responses and aid passenger navigation.
Multilingual support
The NLP engine detects the input language and responds in kind, serving Dubai Airport's highly diverse, international passenger base with fluency and cultural appropriateness.
Psychological AI
An emotional-intelligence module reads sentiment and adapts tone — diffusing frustration, building rapport, and tailoring recommendations to each passenger's needs and travel context.
Projected impact
Figures are projections from initial pilots and simulations. The system is in active research and not yet commercially deployed.
Projected increase in passenger satisfaction with AI assistance
Faster average query resolution
Projected lift in retail & dining revenue
Reduction in customer-service staff workload
“Our experience with Nucleus AI and Raakin was great, a lot of innovative, excitement, and futuristic view on the product and how it will look in the future.”
Build your own with Nucleus
We design custom AI infrastructure for organizations with complex, high-stakes knowledge. Tell us what you're solving.