Three AI systems inside one multi-billion-euro aviation group

The context

Our client is a multi-billion-euro global aviation group — an enterprise environment where every AI system has to satisfy IT, legal and the business at once. Under NDA, Sid Ali Temkit, PhD, personally delivered three production AI systems for the group: a cross-document RAG system, an internal AI framework its teams build on, and email-to-quote automation.

The AI we built

Three systems, one foundation. Each of the use cases below runs in production inside the group — built on the same consistent, EU-hosted architecture, so every new AI use case starts from a standard the group already trusts.

Cross-document RAG

Teams inside the group regularly need to understand how two documents relate: what one says, what the other says, and exactly where they differ. Done by hand, that means reading both end to end and holding every difference in your head.

We built a RAG system that compares two documents side by side. Users ask questions across both at once, and every difference and every answer is grounded in the exact passages it comes from — with citations, so each answer can be checked against the source.

Ask questions across two documents at once, in plain language.
Differences and answers grounded in the exact passages — never a summary from memory.
Citations on every answer, so nothing has to be taken on faith.

An internal AI framework

A group of this size doesn't need one AI tool — it needs a way to roll out many, without reinventing architecture, hosting and governance every time a team has an idea.

We built the group an internal AI framework: a standard, reusable foundation its teams build on. Every new AI use case inside the company starts from the same consistent architecture — EU-hosted and governance-ready — instead of from a blank page.

One consistent architecture for AI use cases across the group.
EU-hosted and governance-ready by design.
New use cases reuse the foundation instead of starting from scratch.

Email-to-quote

Quote requests arrive as free-form emails. Every sender phrases things differently, and someone has to read each message and re-type the details before a quote can go out.

We built AI that reads inbound quote-request emails and turns them into structured, ready-to-review quotes: the unstructured email becomes structured commercial data, and that data becomes a draft quote a human reviews and sends.

Unstructured email in, structured commercial data out.
A draft quote ready for human review — not an auto-sent reply.
Faster quote turnaround, less manual re-typing.

Hosting: EU-hosted, GDPR-compliant — the same standard as everything we ship.

The result

The group now runs three production AI systems on one EU-hosted foundation: document comparisons are grounded and cited instead of manual, inbound quote emails become structured draft quotes, and every future AI use case starts from a framework the group already owns.

Want something like this?

Book a free 30-min call