The decision intelligence platform for airline operations — providing real-time decision support and enabling teams to train, benchmark, and evidence disruption decisions under real-world constraints across live and simulated operations.
"The more you depend on automated systems, the more dangerous it is when those systems fail — and the less capable your people are of taking back control." The case for decision training — BA IT Outage, May 2017
Every incumbent vendor — Sabre, Amadeus, Lufthansa Systems — is building tools that make operations faster at optimising within a disruption. They are not building tools that make operations teams better at deciding when those optimisations hit their limits.
The pilot simulator analogy is exact: ATC candidates complete mandatory simulation before touching live traffic. Signallers train on simulators before managing train movements. Ops teams are the only comparable high-consequence role without mandatory simulation.
Shepherd closes that gap. Not as a training product — as a decision improvement system with three deployment modes and measurable outcomes.
Shepherd operates before disruption, during the critical window, and after every event. Each mode is independently deployable. Together they form a decision improvement flywheel that compounds over time.
Scored, scenario-based exercises that build the decision muscle memory that no optimisation tool can substitute. Ops teams train on the exact constraint collisions — crew FTL, slot windows, connection clusters — that matter most when it's real.
The Decision Intelligence Copilot surfaces ranked recovery options with cost and legal exposure before the window closes. The Network Fragility Heatmap shows cascade risk before it materialises. T-656/24 compliance flagging prevents voluntary decisions from voiding the EU261 defence.
Every live disruption event automatically seeds the scenario library. The Post-Disruption Review module replays what happened against the optimal path — converting institutional knowledge into structured, scored training. The gap between what your team did and what they should have done becomes permanent learning, not a folder that gets filed.
Every scenario is timed, cascading, and scored. Ops teams face the exact constraint collisions that produce bad decisions in live operations — crew FTL, connection clusters, and slot windows simultaneously.
The Decision Intelligence Copilot doesn't replace the ops team's decision. It ensures the decision is made with complete information — cascade impact, cost exposure, legal liability, and codeshare passenger visibility — before the window closes, not after the cascade has started.
Scoring is system-derived from pre-tagged scenario options. Deterministic, consistent, and benchmarked against industry norms — not against a facilitator's subjective assessment. Every session produces the same quality of evidence regardless of who runs it.
Shepherd satisfies all eight IOSA DSP training evidence requirements in a single platform. Every session generates structured evidence automatically — no parallel documentation system required. The 450 airlines on a two-year IOSA renewal cycle are entering audit preparation continuously.
Unlike point solutions, Shepherd compounds. Historical events seed scenarios. Scenarios produce decisions. Decisions produce norming data. Norming data improves assessment accuracy. Assessment accuracy attracts more sessions.
Shepherd was founded by a former British Airways disruption product owner — responsible for the tools, processes, and decisions governing how thousands of disrupted passengers were managed at Heathrow every day.
The May 2017 BA global IT outage — one of the most significant operational events in the airline's history — is where the decision gap became undeniable. Shepherd was conceived inside the disruption events it is designed to prevent.
The founding team's predecessor disruption intelligence project was selected for the IATA accelerate@iata programme — a competitive cohort of the highest-potential aviation startups globally. Founder-level validation of the problem space.
Category signal: Navi AI's $6.7M raise from United Airlines Ventures confirmed investor conviction in AI-powered operations intelligence. Navi operates at the pilot skills layer. Shepherd owns the operations decision layer Navi doesn't touch.
Request a briefing. We'll walk you through the platform, run a sample scenario with your team, and map Shepherd against your specific IOSA audit cycle and disruption profile.