Riyadh Air and IBM Team Up to Build an AI-Native Airline
Riyadh Air is taking a different path: bake AI into operations before scale, not bolt it on later. Through a three-year partnership, IBM Consulting is helping the airline move from pilots to production as it targets its first full commercial service in early 2026.
Initial flights have started. The next phase focuses on a personalized digital workplace powered by AI agents to speed up decisions, streamline work, and smooth the passenger experience.
What "AI-native" means for operations
AI is treated as core infrastructure, not a feature. Processes, data models, and decision flows are being set up so AI agents can assist staff from day one.
The digital workplace will serve role-specific insights and suggested actions. That means fewer tool hops, clearer priorities, and faster response during irregular operations.
Why this matters for ops leaders
- Greenfield advantage: Set data standards, access controls, and workflow ownership early. Prevent future rework and policy sprawl.
- Decision speed: AI agents can surface options with context so teams move from gathering to acting. Human oversight stays in the loop.
- Consistency at scale: Playbooks encoded into agents reduce variance across shifts, stations, and partners.
- Passenger impact: Better staff tools typically show up as fewer handoffs, clearer updates, and tighter turn times.
Execution window: now through 2026
The timeline is tight, which puts a premium on sequencing. Here's a practical rollout lens you can apply in any large operation:
- Data foundation: Map critical systems, event streams, and master data. Define a single "source of truth" for schedules, crew, maintenance, and customer events.
- Agent design: Start with narrow, high-frequency decisions where you have clean data and clear guardrails (approvals, thresholds, audit logs).
- Pilots in live ops: Deploy to one station or team with clear success criteria and shadow mode before full activation.
- Controls and safety: Implement versioning, monitoring, fallback procedures, and post-incident reviews for every agent in production.
Early AI use cases to expect
- Personalized work queues: Role-based dashboards that rank issues by impact and urgency.
- Decision support: Suggested actions with rationale and predicted outcomes, routed to the right approver.
- Knowledge assistance: Fast answers from SOPs, manuals, and historical ops data with citations for audit.
Note: These are proven patterns across airlines and complex operations; the partnership's initial feature set centers on a personalized digital workplace with AI agents.
Risks to manage up front
- Data quality and drift: Poor inputs lead to noisy suggestions. Track feature quality and retraining cadence.
- Safety and auditability: Keep human-in-the-loop on safety-critical decisions with clear logs and signoffs.
- Latency and reliability: Agents must respond within operational time windows and degrade gracefully.
- Change fatigue: Train for new workflows, not just new tools. Measure adoption and confidence, not only uptime.
- Vendor lock-in: Use open interfaces and a data contract so you can swap components without re-plumbing.
KPIs that actually move the needle
- Decision cycle time: From alert to action, by scenario.
- Right-first-time rate: Actions that require no rework or escalation.
- Agent adoption: Daily active users and task coverage by role.
- Passenger impacts: Missed connections avoided, on-time performance, queue time, NPS deltas tied to agent use.
- Safety and compliance: Exception rates and audit findings related to AI-assisted actions.
What to do next (even if you're not greenfield)
- Stand up a minimal digital workplace that consolidates alerts, runbooks, and approvals in one view.
- Pick one high-value decision loop and add AI suggestions in shadow mode. Compare outcomes before flipping to assist mode.
- Create a lightweight governance pack: data owners, model owners, guardrails, rollback plan, and weekly review.
- Upskill frontline leaders on prompting, validation, and exception handling so agents become teammates, not tickets.
Riyadh Air's bet puts a spotlight on AI-driven operations through 2026. The lesson for every ops team: build the groundwork early so scale doesn't amplify gaps.
Learn more
IBM Consulting * Riyadh Air
Skill up your team
If you're building similar capabilities, see role-based options at Complete AI Training: Courses by job or the AI Automation Certification.
Your membership also unlocks:
