Karthic Sundararaj, VP Technology and Architecture of Etihad Airlines

Summary

• Etihad Airways, the UAE national carrier onboarding 23 new aircraft this year as part of aggressive growth through 2030, has unified its data platform on Databricks since 2021 and is now scaling AI-powered self-service analytics across finance, operations, and customer experience.
• This video shows how Etihad's CFO uses AI/BI and Genie to run natural-language queries — including a now-famous 'orange juice story' that illustrates procurement cost optimization — without needing to involve the data team.
• Etihad treats the Databricks Data and AI platform as an operating system for AI, connecting Genie spaces across NPS, on-time performance, and financial data, with Unity Catalog providing the governance foundation that ensures accurate, trustworthy answers.

Karthic Sundararaj, VP Technology and Architecture of Etihad Airlines

Hear from Etihad Airlines' VP Technology and Architecture, Karthic Sundararaj, on how they use Databricks and Genie.
Talk By: Amber Roberts, Staff Tech Marketing Engineer, Databricks ; Ari Kaplan, Global Head of Evangelism, Databricks ; Karthic Sundararaj, VP Technology and Architecture , Etihad Airways ;

Chapters

• 00:00 Introduction: Etihad Airways at the Databricks Data and AI Summit
• 01:09 Etihad's growth: 23 new aircraft and fastest-growing airline trajectory
• 02:01 Karthic Sundararaj's role: technology, engineering, architecture, and data
• 02:49 Data journey: unified platform since 2021 and scaling data products
• 03:28 CFO self-service with AI/BI and Genie for cost and supplier analysis
• 04:42 The orange juice story: natural-language procurement optimization via Genie
• 05:09 Connecting Genie spaces: NPS, on-time performance, and cross-domain analysis
• 06:23 Unity Catalog as the governance foundation for accurate AI answers
• 06:47 Genie Ontology and Databricks as an operating system for AI

FAQs

How does Etihad Airways use the Databricks Data and AI platform?

Etihad Airways has been on the Databricks Data and AI platform since 2021, using it as the foundation for a unified data platform that now powers self-service analytics, AI/BI dashboards, Genie spaces, and AI-driven insights across finance, operations, and customer experience. The airline treats Databricks as an operating system for AI as it scales toward 2030.

What is the Etihad orange juice story?

The 'orange juice story' refers to how Etihad's CFO used Genie — Databricks' natural-language analytics interface — to investigate procurement costs for orange juice across suppliers and routes without needing to involve the data team. It became a compelling internal example of how self-service AI analytics can deliver direct business value to non-technical executives.

How does Etihad use Genie for cross-domain analytics?

Etihad connects multiple Genie spaces covering NPS (customer satisfaction), on-time performance, and financial data, allowing analysts and executives to ask cross-domain questions that previously would have required multiple teams. Unity Catalog acts as the governance layer that ensures those answers are accurate and grounded in trusted, well-described data.

What role does Unity Catalog play in Etihad's AI strategy?

Unity Catalog is Etihad's governance foundation, providing the semantic layer and data descriptions that allow Genie to return accurate, business-contextualized answers rather than raw query results. In this video, Karthic Sundararaj explains that the quality of Genie's outputs depends directly on how well the underlying data assets are catalogued and described in Unity Catalog.

Full transcript

[00:21] Hey, welcome back everybody. I'm Ari Kaplan, one of the co-hosts of Summit Live and I am very delighted to have actually the very first company from the United Arab Emirates who's ever been not just on Summit Live but presenting at
[00:33] Summit, Karthik who's the CTO of ETA. A big warm welcome. >> Thank you, Ari. and uh thanks for having me. It's a pleasure being in uh San Francisco. It's my first uh summit and I
[00:45] am very impressed with the scale. I really appreciate it. >> How do you like it so far? >> Oh, it's been incredible. It's been incredible. I have enjoyed uh your announcements. Uh so we just need to see how it translates into our business now.
[00:57] But uh very excited. >> Absolutely. So first start like uh I I know a lot about Edahad, but we have 160 countries here. Uh tell us about the airline. >> Yeah, sure. Sure. Uh we are the national
[01:09] carrier of United Arab Emirates. We are based in Abu Dhabi. Um Etihad uh is is is growing at an incredible pace. Uh we we just recently and I launched and uh
[01:23] opened up four new destinations and we are one of the fastest growing airlines in the world. We are on boarding 23 aircrafts this year and that continues all the way to 2030. So an aggressive growth plan and a fantastic airline. if
[01:35] you haven't flown one. >> I uh everyone should try flying. I personally have many times. Um we were talking about like you're a hub for many many destinations. So when I fly like for example from India through um Abu
[01:49] Dhabi on Edah and then back to the states on on uh other airlines as well. Um so um and I also want to hear about your role and your team, your CTO. What does that look like?
[02:01] >> Sure. Absolutely. Um so I'm responsible for uh technology within Etihad. So that includes engineering, architecture, platform. Uh I also have a small lean agile center of excellence and more
[02:13] importantly I run the data team as well. Uh which we will explore a little bit here. >> Yeah, absolutely. Um the first thing I want to say is so I was there just a few months ago. I think it was in February
[02:25] and had the the honor to be welcomed by you. uh beautiful offices. Um and I remember a giant model airplane out front, great cafe, but then I had the
[02:37] opportunity to meet with some of the other divisions like their uh chief financial officer and and the control room and like the operations was very much impressive. But you were starting
[02:49] to talk about the data. What does your data operations look like? Yeah. So data is incredibly important for Ethihad. We the journey started in uh in about 2021
[03:01] um just after I joined. So we focused a lot on consolidating our data platform. Um that was step nine. We now have a single unified data platform and then we started to build data products on top
[03:14] and now we are launching self-service. So uh the CFO story was a great success because this he has self-service on um on the BI stack. So the AI toBI stack
[03:28] uses Genie to find out a lot about his cost and cost control. I think that's a great success story for Etihad. Now we're slowly scaling our Genie implementation to all the other
[03:40] departments and uh we'll see where where it t this journey takes us but we also try to look at uh the products and services that you have announced now to implement that coming >> that's great and I'm not saying it just
[03:52] since you're a guest now but you are one of the premier examples I have of Genie when I travel and talk to customers since you're saying the CFO and really the his broader team using Genie you
[04:04] have to have the data governed um and and you know the semantic layer controlled. So when someone asks like a research question like how is our supply chain or uh should the what if the price
[04:17] of oil goes up or down how should we arbitrage if a CFO you know trusts in that governed through your department then you know really it just points to you know the the the governance of unity
[04:30] catalog. Yes, absolutely. It's uh the Unity catalog has been a gamecher for us and we are now scaling Unity catalog across. Uh we have the famous orange juice story, you know. Um there was a
[04:42] there was a catering question which is where the implementation of Genie and self-service started. Um and and it was uh interesting for not only the the the VP of procurement but also the CFO to
[04:55] write specific questions and then optimize our uh supplier costs uh which started with that now we have fuel we have other other genie spaces and we're now super excited to connect all these
[05:09] genie spaces together. So an an example would be uh how would the NPS be affected with uh delay in on-time performance or what is the relationship.
[05:21] So it's bringing these different data sets but also bringing the the element of orchestration into it as well >> and NPS is the net promoter score like sentiment of your >> that is correct that's that's a
[05:34] measurement of how successful we are or how happy our customers are. So yes, >> and that's great. I remember you that was the the two topics. Well, you started saying the orange juice story and we wrote a blog on that. It's just
[05:46] you can ask some open-ended questions like what vendor is most efficient or inefficient and our supply chain and it uh Genie with your data and with your you know business team you know guiding
[05:58] you with the guidance to make sure it's accurate could tell you what vendors are uh you know o overpaying underpaying. >> Yeah. Exactly. Exactly. the the the context uh um problem was an interesting
[06:10] problem we couldn't solve before but now with uh with the unity catalog and the the context Genie can uh get grounded answers and do some research and it's
[06:23] very impressive. So, so hence I am also very excited about the the the ontology layer that you've just announced now to see how um how we can scale with that
[06:35] you know truly impressive in the sense that uh you maybe I coin a term here it you're really building an operating system for AI with the data backbone. So let's see let's see where we go with
[06:47] this but very exciting. >> Yeah very good. So where you go with it? You've heard some announcements like uh are there other use cases you're working on beyond finance or other products
[06:59] you're >> Yes, I think we are scaling Genie all the way through. Um we are interested in uh a part of the CDP um and then the rest is uh to to to look at Genie Genie
[07:12] 1 as a as a core platform. So yes, absolutely. >> Well, I think you'll like it. I know it was kind of announced yesterday not just on your mobile phone but on your laptop but every almost everyone at data bicks
[07:24] is already using it uh you know dashboard but just asking more forwardlooking research type of questions. So um any uh anything else you wanted to part with? >> No I think I think it's uh the vision is
[07:36] very good. We are a great uh great partnership going and uh super excited for the future. Thank you very much. >> Thank you for coming on board. It's been a pleasure. >> Thank you for hosting me. Thank you.

Learn more about the Databricks Data and AI platform.