OpenAI has the
aviation sector sewn up. Or so it would seem from the raft of announcements the
artificial intelligence (AI) platform has made this year.
In April, PhocusWire
covered Singapore Airlines’ partnership with the platform, which was established
to help with customer service, operation efficiency and staff productivity.
In July, Air New
Zealand announced a deal to
give its 3,500 staff access to ChatGPT Enterprise to expand AI use
across the airline and improve efficiency and customer outcomes. In November,
Emirates announced it would
deploy ChatGPT Enterprise across the group to
“tackle complex commercial challenges, strengthen operations and enhance the
customer experience.”
In part, this is
because of the first-mover advantage the platform gained after it brought large
language models to the world’s attention
in November 2022.
This has translated
into widespread adoption by enterprises. In early November, OpenAI announced
that it had more than 1 million
business customers directly using its
services, leading to the company to claim it is “the fastest-growing business
platform in history.”
Virgin Atlantic’s
strategy
Virgin Atlantic is another
airline that has chosen OpenAI as its AI platform of choice.
In a recent interview, Virgin Atlantic CFO
Oliver Byers said, “When we started exploring enterprise AI a few years ago, we
took a broad approach—testing, learning and seeing where the real value lay.
Over time, we narrowed our partnerships and went deep with a few select
leaders, including OpenAI. The conviction came quickly: We could see tangible
benefits day to day, from early pilots to our ChatGPT Enterprise trials.”
Virgin Atlantic is
using OpenAI’s Codex agent.
“We’re writing and
testing code faster, shipping features more quickly and improving customer
experience at speed,” Byers aid.
Its most recent
AI-powered launch is its “Concierge” digital assistant.
Speaking to PhocusWire,
Virgin Atlantic’s head of digital product and journeys, Nathan Bolt, said the
ChatGPT-powered concierge is the “next best thing to carrying a little Virgin
Atlantic employee around with you 24-7.”
“We wanted to
replicate how our people talk to our customers, so there is a bit of
playfulness where appropriate in the experience,” he said. “As part of the
project, we interviewed a lot of our colleagues to understand how they interact
with people and then injected that into the prompting.”
Concierge will assist
Virgin Atlantic passengers with four use cases: holiday planning, booking a flight, questions about Flying Club points and tiers and general inquiries, such as the
number of bags allowed or which movies are being shown onboard.
In 2026, the airline
will add real-time voice to Concierge.
“At the moment, it's
a text-to-speech experience, so you can talk to it and it does talk to you, but
that isn't quite as natural as the real-time voice experience,” said Bolt. This
will help for hands-free phone usage or when English is not the user’s first
language.
Bolt said the airline
is also looking at integrating agentic flows within Concierge where
appropriate.
“You can change your
booking online today through a traditional website experience. We could enable
that through Concierge if the value is there in doing it in an agentic way. We
might start to augment some of our experiences with AI assistance within that
core experience rather than having it as two separate flows.”
Concierge will also
be at the heart of the airline’s new app which will be launched “quite soon,”
he said.
Additional AI applications
More generally, AI is
creeping into airline tech stacks by stealth rather than direct deals with
platforms.
Bert Craven, deputy
CEO of airline tech procurement consultancy T2RL, told PhocusWire that most
airlines are not running AI platform tenders.
“What's actually happening is
that the providers of existing services that our clients have are adding AI to
their products. The airline client is then coming to us and asking how to
contract for these additional services, and a lot of this is around risk
mitigation.”
Ian Tunnacliffe, T2RL’s SVP
of consulting, added, “Outside a small group of very large airlines, most
airlines do not have the capability of managing all of this complexity in-house,
so there is a big opportunity for the systems integrators of the world like
Infosys and TCS.”
According to Craven, people
are seeing AI benefits in “more efficient processes, taking people out, cutting
costs, increasing efficiency, that kind of stuff.”
“Whether or not that turns
out to be true is going to play out on a case-by-case basis. At the moment,
everyone is saying the entire universe is possible. What it may turn out is
that AI is not the silver bullet that everyone thinks it is.”
Craven said that using AI
to make operations more efficient could provide a return on investment—“all
these things where a tiny percentage change in the trim of an aircraft, for
example, can save enough fuel to be the difference between a profitable airline
and an unprofitable airline.”
Aviation tech specialist SITA
is also seeing this trend.
Simon Lomas, SITA’s CIO
of technology and engineering, told PhocusWire that predictive maintenance, crew optimization and
passenger experience enhancement are popular use cases for AI.
“Airports
are also scaling AI across baggage handling, security screening and passenger
flow management, while generative AI is reshaping customer engagement through
personalization and automated service interactions,” he said.
Lomas said
the ROI of AI is very much front of mind,
particularly as AI companies try to recoup the vast investments they have made
in developing their platforms.
Some
analysts have pointed to reports that OpenAI is losing $3
for every dollar it makes. It is certain that other platforms with smaller user bases are losing
even more.
“While
significant investment has gone into developing AI capabilities, the aviation
industry is focused on ensuring any adoption is driven by clear business value
rather than technology for its own sake,” said Lomas.
“Airlines
and airports are looking closely at use cases that deliver measurable
efficiency gains, cost savings or revenue impact, which helps maintain a
balanced relationship between investment and long-term value.”
T2RL’s Craven
said this may not be an issue if airlines are using AI for short-term proofs of
concept, but he also offered a warning.
“Today's
innovation is tomorrow's obligation. If innovative services are later withdrawn
because the costs become unsustainable, consumers will blame the airline, not
the technology provider. If you go out with something that's really whizzy that
consumers love and then you have to yank it away because it's now costing $50 a
user, that is going to be a very painful removal.”