Aviation is among the most data-intensive industries on the planet. Each flight generates a torrent of data: gas consumption, engine telemetry, passenger preferences, real-time climate patterns and extra. For Aer Lingus, Eire’s flagship provider, this complexity is compounded by a storied historical past. Many airways nonetheless function on methods constructed many years in the past, the place knowledge is trapped in departmental silos. On this atmosphere, making easy selections can require handbook knowledge extraction and weeks of research.
Dave O’Donovan, Chief Digital, Information & Transformation Officer, Aer Lingus, is main this cost. Below his management, Aer Lingus has undergone a radical shift, redirecting a good portion of its capital spend away from conventional IT upkeep and towards a unified platform powered by Databricks.
I sat down with Dave to debate the mechanics of this transformation. We explored how Aer Lingus is transferring previous legacy to a totally digitally led buyer expertise, and why he believes the key to AI success is knowledge literacy.
Shifting infrastructure spend to the information basis
Aly McGue: Aer Lingus is 90 years outdated. That’s an unimaginable milestone, but it surely additionally comes with the problem of legacy methods and processes. How are you framing the corporate’s mission as we speak within the context of a quickly evolving digital panorama?
Dave O’Donovan: It’s an enchanting time for us. Aer Lingus is Eire’s window to the world. We now have an enormous short-haul community throughout Europe, and we’re truly the second-largest European provider on the North Atlantic by US locations served. However being 90 years outdated means now we have methods and mindsets which have matured over many years.
Our mission now’s to keep up that well-known “heat welcome” and caring model identification whereas assembly the expectations of a traveler who’s extra digitally savvy than ever and needs premium experiences. That forces us to ask: How do we provide a self-service, digital-first expertise that also seems like Aer Lingus? The reply, invariably, is knowledge.
Aly: You’ve made a really daring transfer not too long ago by redirecting a sizeable proportion of your IT and alter spend particularly towards knowledge. What led to that “all-in” second?
Dave: It was a collective determination on the administration committee degree about 18 months in the past. We reached a degree the place we realized that understanding leverage AI is not a “good to have.”
For years, many corporations, together with airways, may get away with under-utilizing their knowledge. However the tempo of AI evolution has been like gasoline on a hearth. We determined that as an alternative of chasing each new “shiny factor” our opponents introduced, we might cease and lay the foundations. We’ve spent the final yr and a half targeted on the platform, governance, knowledge high quality and, most significantly, knowledge literacy. If you do not have these stable foundations, any AI you construct is only a home of playing cards.
Aly: Many organizations battle with the transition from legacy knowledge warehouses to a contemporary structure. How did your place to begin at Aer Lingus affect your option to go along with Databricks?
Dave: Surprisingly, we felt fortunate that we had been a bit slower to maneuver than a few of our friends. We hadn’t made huge investments within the “first wave” of cloud knowledge instruments, so we did not have to fret about writing off latest sunk prices. We nonetheless had many legacy on-premises warehouses.
After we seemed on the market, it had matured. It was clear that Databricks provided a “soup to nuts” resolution. We may go all-in on a single lakehouse structure. What actually clinched it for me wasn’t simply the suggestions from our knowledge engineers — who beloved the efficiency — however the imaginative and prescient for democratizing knowledge. I’m enthusiastic about issues like Databricks’ knowledge warehousing platform and Databricks Genie. These instruments enable enterprise customers to ask questions of the information in plain English. That’s the solely option to actually scale.
Eliminating the legacy IT bottleneck
Aly: You talked about the “bottleneck” of legacy methods. In the event you may snap your fingers and take away one impediment between your knowledge and a ultimate determination, what would it not be?
Dave: It will be the bodily extraction of knowledge from methods which are “60 years younger,” as we wish to say. These legacy methods are implausible at what they had been constructed to do — working an airline safely — however they weren’t constructed for the age of generative AI.
We have to transfer from a world the place a division says, “That is my knowledge, I personal it,” to a world the place knowledge is a shared, holistic asset used to enhance your complete operation.
Aly: Let’s speak about that human ingredient. You’ve invested closely in a “Information Literacy Academy.” Why is that such a precedence for an airline government?
Dave: As a result of instruments are solely half the battle. You’ll be able to have the most effective LLM or the quickest compute on this planet, but when your groups haven’t got the instinct or the abilities to make use of them, you’ve gained nothing.
We partnered with a UK-based group to construct a customized curriculum. We’ve finished every part: on-line coaching, in-person workshops, and even recording our personal podcasts. However even with all that, you must push it each single day. It must be top-down. Our CEO is consistently encouraging groups to consider knowledge literacy. We attempt to present “bite-sized” chunks of data that folks can use of their day jobs instantly.
My aim is that, in 5 years, “citizen builders” would be the norm at Aer Lingus. If we nonetheless have a state of affairs the place a enterprise chief does not know exploit knowledge to run their division, then I’ve failed in my function.
The aggressive benefit of real-time insights
Aly: In an trade like aviation, “real-time” is a requirement. The place are you seeing the largest affect of real-time insights as we speak?
Dave: The Operation Management Heart (OCC) is the guts of the airline. About 24 hours out from a flight, the variables begin transferring quick: climate patterns change, crew availability shifts and plane upkeep points may pop up.
Previously, these selections had been usually made in silos. Now, by pulling knowledge from numerous sensors throughout the operation right into a unified platform, our OCC groups can see the “full image” in actual time. If now we have to cancel a flight or take a delay, we would like that call to be based mostly on essentially the most present knowledge doable to reduce disruption for our prospects.
On the industrial aspect, it’s simply as important. We promote over 80% of our tickets by direct digital channels. We’re a high-volume retail platform. Having the ability to use real-time insights to regulate pricing — making certain we maximize our load whereas additionally maximizing yield — is an enormous aggressive benefit.
Modernizing with agentic AI
Aly: How are you experimenting with AI brokers as we speak? Do you’ve gotten a particular use case in thoughts?
Dave: We’re beginning with one thing “good and easy” however extremely frequent: enterprise case growth. In any massive group, you spend an enormous period of time writing enterprise circumstances to get funding.
We’re taking a look at an agentic workflow the place an agent helps you craft the case. Then, we would like a “CFO agent” to overview the case and establish precisely what the CFO will ask. It’s a good way to stress-test our inside logic earlier than we ever even step into the assembly room.
Aly: With the tempo of change being so quick, how do you steadiness that pressing have to “scale now” with the truth of experimentation?
Dave: It’s a fragile steadiness. It’s very simple to get distracted by “shiny issues” to maintain your board or CEO glad within the quick time period. However you possibly can’t lock your self in a closet for 18 months to construct the “good” platform both.
I observe a 75/25 rule. About 75% of our capability is targeted on the long-term foundational technique — getting the information high quality and Unity Catalog governance proper. The opposite 25% is targeted on innovation and fast market worth development. You want these small wins to keep up momentum and hold the enterprise engaged. We even arrange a devoted “Steady Enchancment” group of about 20 individuals who go round to completely different departments — finance, buyer care, operations — and redefine processes so they’re “AI-ready.”
Constructing a pivot-ready tradition to scale AI
Aly: Lastly, what’s your recommendation to different CDIOs who really feel the stress of this AI hype cycle?
Dave: Do not deal with being “future-proof,” as a result of you possibly can’t be. The expertise adjustments each six to 12 months. As a substitute, deal with being “pivot-ready.”
Accomplice with platforms like Databricks which are constructed on open requirements and open supply. That provides you the pliability to vary path because the market evolves. And most significantly, spend money on your individuals. Essentially the most beneficial individuals in my group are these with curiosity, instinct and creativity. In an period the place expertise is changing into commoditized, these human qualities are your solely true aggressive benefit.
Closing Ideas
Dave’s strategy at Aer Lingus serves as a masterclass in fashionable digital management. Whereas the trade fixates on the generative potential of AI, he has targeted his mandate on the one variable that determines a corporation’s final ceiling: its individuals.
By treating knowledge literacy as a business-wide crucial slightly than a technical elective, Aer Lingus is fixing the elemental problem of the AI period. They are not simply modernizing a legacy airline; they’re constructing a resilient, data-fluent tradition the place each worker is provided to show uncooked data into operational excellence, in a sector the place seconds matter in decision-making. That cultural basis is the final word aggressive moat.
To find how greater than 25 trade specialists are charting a course towards profitable AI deployment, entry the “Making AI Ship” report from Economist Enterprise, produced with assist from Databricks.
Watch the complete interview with Dave O’Donovan beneath