Engineered intelligence,
not just applications.
Three complex business decisions at Fly91 — where Algen engineered intelligence into the process itself.
Fly91 is an Indian regional airline. Algen embedded with their team and built three intelligence layers across operations, data access, and revenue management — each designed around the DIRECT method.
From dashboards to decision intelligence
Same data, different roles, different questions. A revenue manager, an operations lead, and a station manager all need different answers from the same operational database.
The Problem
Every role was looking at the same dashboards but asking different questions. Teams maintained parallel, manually-updated reports just to get the view they needed.
What Algen Built
A role-aware intelligence layer that understands who is asking and surfaces the insight relevant to that role — not a generic export for everyone.
The Workflow
The Outcome
One data layer now serves every role with the specific view each one needs.
Eliminated the parallel, manually-maintained reports each team relied on before. The same underlying data, interpreted differently for each decision-maker.
ENGINEERING NOTE
Built as a semantic layer over Fly91's existing operational data — not a replacement for the dashboards already in place, but an interpretation layer in front of them that adapts based on identity and role.
Making enterprise data conversational
The data and the dashboards exist. What's missing is a way to ask a plain question and trust the answer.
The Problem
Business users had dashboards but not the query skills to use them. Every non-trivial question required a data team request and a multi-day wait.
What Algen Built
A controlled natural-language interface that lets business users ask questions across approved enterprise data sources — with validation before any answer returns.
The Workflow
The Outcome
Business users get trustworthy answers without writing a query.
Questions that previously took days of back-and-forth with the data team now get answered in seconds — with the same data integrity guarantees.
ENGINEERING NOTE
Query generation grounded in a semantic model of the schema. A validation layer catches wrong or unsafe queries before they run. Access scoped strictly to approved data sources.
Turning demand signals into pricing decisions
Fare-class allocation depends on signals that move independently. No single number tells you the right answer.
The Problem
Revenue managers allocated fare classes using manual judgment and spreadsheets. The signals they needed — booking velocity, yield, competitive pricing — were scattered across systems.
What Algen Built
A revenue intelligence engine that evaluates multiple demand signals together and recommends fare-class allocations — with the human making the final call.
The Workflow
The Outcome
Data-grounded fare recommendations — with the human still deciding.
Revenue managers now see an explainable recommendation instead of relying on manual judgment alone. They can see why, trust it, and override it when their experience says otherwise.
ENGINEERING NOTE
A model architecture combining historical and current operational signals into explainable recommendations. Integrated directly into Fly91's existing revenue-management workflow.
The Fourth Rung
See. Ask. Decide. Control.
Every system above depends on agents acting inside real enterprise data. Traccia is the control plane Algen builds in — so what these systems do stays observable, evaluable, and governed.