How Yepic and Oracle Made Complex Aviation Data Speak
Bringing clarity to operational complexity
The primary directive was to help authorised users understand where aircraft were, what assets were available on them and what the wider operational picture meant. The difficulty was not a shortage of data. It was that the information lived across complex systems, served people with very different responsibilities and included material that could only be disclosed at the correct level of security clearance.
A single generic dashboard or chatbot would not have been enough. The system needed to find the right information, respect permission boundaries, interpret the result and communicate it at the right depth for the person asking.
One body of operational data. Different answers for different authorised users.
A management user might need a concise view of airport holdings and fleet activity. An operational user could ask how many aircraft of a particular type had been dispatched. With the appropriate clearance, a specialist could query more detailed information about an aircraft and what it carried, including luggage and restricted or military-linked assets.
The public examples are deliberately high-level. No live operational data, database structure or restricted specification is disclosed in this case study.
The Oracle-connected query system
Working with Oracle, Yepic created a natural-language query layer capable of reaching across Abu Dhabi Aviation’s databases. A user could ask a management or operational question conversationally instead of navigating several systems, reports and data tables.
From raw results to an interactive explanation
Returning a database value was only the beginning. Yepic built an interpretation layer that turned queried information into a series of interactive charts and graphs. The avatar then narrated those visualisations, explained what was significant and made the data accessible to people with different levels of technical understanding.
This was the avatar’s superpower. It did not sit beside a dashboard as decoration. It became the conversational interface to the intelligence system: asking for clarification, presenting an appropriate visual answer, explaining trends or exceptions and allowing the user to continue with follow-up questions.
The same system, three levels of understanding
- Management: concise answers about airport holdings, fleet position and operational performance, supported by clear visual summaries.
- Operational teams: more detailed questions about dispatch activity, aircraft types, availability and the assets assigned to particular aircraft.
- Highly cleared specialists: deeper interrogation of restricted operational information, including aircraft contents and sensitive asset categories.
The information changed not only according to what a user was allowed to see, but also according to how much explanation they needed. The aim was to make the same complex environment useful to a senior decision-maker, an operational colleague and a technical specialist without overwhelming one user or under-informing another.
A demanding first use case
Abu Dhabi Aviation also asked whether Yepic could create avatar-led aircraft safety content in multiple natural languages, including the visual requirement to demonstrate safety equipment. The client’s stated implementation window was three weeks.
What Yepic delivered
- A natural-language query experience connected with Oracle enterprise data systems.
- Role- and clearance-aware access to different tiers of information.
- An interpretation layer that adapted answers to each user’s level of understanding.
- Interactive charts and graphs generated from query results.
- A dedicated real-time avatar that narrated the data and supported follow-up questions.
- Dedicated development and production environments.
- API keys, endpoints, streaming and iframe integration.
- Captions, microphone behaviour and interface controls.
- Latency, WebRTC, browser, mobile and corporate-network testing.
- Cybersecurity, data-protection, go-live and ongoing production support.
The outcome
The programme progressed into operational use and remained important after go-live. Yepic helped turn a complicated, permission-sensitive data environment into something people could question, see and hear — without reducing every user to the same dashboard or exposing restricted information outside the correct clearance level.
Why it matters
This project demonstrates a broader enterprise pattern: conversational AI is most valuable when it sits on top of governed data, understands who is asking and can convert a complex result into a clear decision aid. The avatar was the visible superpower, but the value came from the complete system behind it — query, authorisation, interpretation, visualisation and human explanation working together.
Why these images are illustrative
Because this project involved government security and privacy requirements, and the aircraft involved was connected to military applications, Yepic cannot publish screenshots, interface captures or detailed implementation architecture from the live deployment.
The images below are AI-generated conceptual visualisations intended only to illustrate the use cases and delivery environment. They are not photographs or screenshots of the deployed system.
The secure aviation intelligence experience, illustrated
