Transforming Travel with AI - Insights from Acai Travel and Amadeus

In this episode of the Travel Tech series, we speak with Riccardo Vittoria, CEO and Co-Founder, Acai Travel, and Gabriel Langella, VP of Customer Operations, Travel Distribution at Amadeus. The conversation explores how AI is Transforming Travel operations, the power of startup-corporate partnerships, and what the future holds for human and Artificial Intelligence (AI) collaboration in the industry. 

Tune in to this episode to understand: 

➤ Why contact centers are becoming the control tower for next-gen travel operations and how Acai Travel is using AI agents and supervisors to automate intent recognition, task execution, and decision-making in real time

➤ The power of startup-corporate partnerships to drive meaningful innovation: hear how collaboration between Acai Travel and Amadeus unlocks speed and scale, combining bold thinking and complementary expertise.

➤ Where AI is going next in travel: from agentless operations and real-time orchestration to human and AI collaboration at scale

What can we expect in this next exciting chapter of our industry? Tune in to find out. 

Speakers: 

• Riccardo Vittoria, CEO and Co-Founder, Acai Travel 
• Gabriel Langella, VP of Customer Operations, Travel Distribution, Amadeus.

To stay up to date on the latest news from Amadeus, and to catch the next episodes in our series, check out the Resource section on our website or follow us on Spotify or Apple podcasts.

Transcript

Elena  0:12
Hi, and welcome to the Amadeus Travel Tech
Talk podcast.

Elena  0:16
Today, we're continuing the Transforming
Travel series, where we take an inside

Elena  0:20
look at how the industry is
on a journey of transformation, 

Elena  0:23
to make the travel work
better for everyone, everywhere.

Elena  0:27
At Amadeus, we believe
that transforming travel

Elena  0:29
means more than just
reaching a destination.

Elena  0:32
It's about making travel effortless,
meaningful, and connected.

Elena  0:36
It's about rethinking how the industry
works, from the systems behind the scenes

Elena  0:41
to the experiences travelers feel.

Elena  0:43
That transformation is powered by four
forces, bold thinking, smart technology, 

Elena  0:49
powerful partnerships, and lasting impact.

Elena  0:51
My name is Elena Ruiz, and I'm an
Investment Manager at Amadeus Ventures, 

Elena  0:55
sitting in our Transversal Corporate
Strategy team, and I'll be your host today.

Elena  1:00
As a quick note, Amadeus Ventures is
constantly looking to strategically invest

Elena  1:04
in technologies and teams across the world
that can help power Amadeus offerings, 

Elena  1:10
while giving, of course, those teams a
front seat to the core of the travel

Elena  1:13
industry, in between both worldwide travel
sellers and providers.

Elena  1:19
But before we dive into today's topic,
let's zoom out for a second, because we're

Elena  1:23
living through one of the most profound
technological shifts in human history.

Elena  1:28
In late 2022, the world
met ChatGPT, and what

Elena  1:31
followed was nothing
short of a global awakening.

Elena  1:35
In just five days, it reached a million
users.

Elena  1:39
Within two months, a hundred million.

Elena  1:42
Today, hundreds of millions of people interact
with Large Language Models (LLMs) daily.

Elena  1:47
from customer service
bots and travel planners

Elena  1:49
to code generators and creative partners.

Elena  1:52
And to me, this is just the beginning.

Elena  1:54
Every few centuries, humanity
makes the leap, the Iron Age,

Elena  1:58
The Industrial Revolution,
the birth of the Internet.

Elena  2:01
Each one has redefined what's possible.

Elena  2:04
Today, we stand at the threshold of the
next great leap, the Age of Intelligence.

Elena  2:10
Generative AI and autonomous agents are
not just tools, they are becoming

Elena  2:13
collaborators, capable
of augmenting human

Elena  2:16
decision-making and
transforming entire industries.

Elena  2:20
What electiity was to factories,
what steam was to locomotion, AI is

Elena  2:24
becoming to knowledge, creativity,
and service.

Elena  2:28
And while headlines often focus on AI's
risks or speculation, this podcast is

Elena  2:34
about the real transformation, where
intelligent systems are already driving

Elena  2:38
performance, efficiency, and customer
experience in industries, like travel.

Elena  2:42
This episode will
give you the context you need

Elena  2:46
to understand where
this revolution is heading.

Elena  2:49
We're starting with a concrete case,
how AI is reimagining contact centers, 

Elena  2:54
not as cost centers, but
as intelligence, always

Elena  2:57
on control towers for
the travel ecosystem.

Elena  3:01
So joining me are Riccardo Vittoria,
CEO and co-founder of Acai Travel,

Elena  3:05
building a next-generation
AI-first operating system

Elena  3:08
for the travel contact
centers. Welcome, Riky.

Riky  3:11
Hi Elena, thank you for having me.

Elena  3:14
And also my dear colleague, Gabriel
Langella, VP of Customer Operations at

Elena  3:18
Amadeus, leading the
teams in charge of bringing

Elena  3:20
AI innovation into the
heart of our operations.

Elena  3:23
Welcome, Gab.

Gabriel  3:23
Hello, Elena.

Elena  3:25
Welcome both, Riky and Gabriel

Elena  3:27
It's fantastic to have you here.

Elena  3:28
I'd like to kick off the session with a
broader understanding of Gabriel's

Elena  3:32
position at Amadeus, because Gabriel,
Amadeus is a huge organization.

Elena  3:35
So can you help us and help
our audience by describing a

Elena  3:38
bit your role and what does
customer operations really mean?

Elena  3:42
Who do you deal with in your day-to-day?

Gabriel  3:43
Yes, it's a pleasure.

Gabriel  3:45
So, I am part of a business unit called
Travel Distribution.

Gabriel  3:51
And that business unit is
essentially dealing with travel

Gabriel  3:54
sellers who distribute
content and travel solutions.

Gabriel  3:57
We interact, of course, with the providers
that are the ones owning that content and

Gabriel  4:01
distributing their content via the supply
chain and via the travel sellers.

Gabriel  4:05
So, in that organization, I am dealing
with operations.

Gabriel  4:08
If I can oversimplify, operations is when
you have signed a contract.

Gabriel  4:12
That contract needs to become a reality.

Gabriel  4:14
This is where we start.

Gabriel  4:16
And then, after, you have to be able to
ensure that your customer can work

Gabriel  4:20
day-to-day to benefit from the value that
the customer expects to get from Amadeus.

Gabriel  4:25
And this is what we guarantee also.

Elena  4:28
All right. So you really sit at the core of
the distribution business within Amadeus.

Elena  4:32
That's great.

Elena  4:34
Now let's dive into what I really is.

Elena  4:36
Because as you almost probably know
already, we made public some months ago

Elena  4:40
that Amadeus Ventures did a minority
investment in Acai back in April this year.

Elena  4:45
And we've already kicked off conversations
for collaboration.

Elena  4:48
But Riky, we will dive a bit deeper into
your background in a minute.

Elena  4:52
But before that, you launched I in 2023.

Elena  4:55
What is I and what's that travel operating
system that we've all heard about?

Elena  5:00
And what's fundamentally broken in today's
travel contact center operations?

Riky  5:04
Hi Elena, Hi, thank you for inviting me.

Riky  5:08
Beautiful to be here.

Riky  5:12
I've been in this industry for a while.

Riky  5:14
I had another company
before, also AI for travel contact

Riky  5:19
centers, which I sold to
American Express, GBT, in 2020.

Riky  5:25
The problem for GBT and all the other TMCs
and OTAs was that demand is growing.

Riky  5:32
Demand came back from COVID.

Riky  5:34
It's growing beautifully
at 5% per year, but

Riky  5:38
everybody's struggling
to service demand.

Riky  5:42
And the more I looked, with I,
the more I looked into different

Riky  5:45
conversations with customers in the OTA,
TMCs and online space, the more I see this

Riky  5:50
pattern that everybody says it's very
difficult to find GDS trained agents.

Riky  5:55
There is a high wage inflation year over
year on agents.

Riky  5:59
Very difficult to train and retain agents.

Riky  6:02
And so the effect of that is that it's
difficult to meet SLAs and it's difficult

Riky  6:08
to provide the season and NPS score and
customer satisfaction.

Riky  6:12
So to address this problem, we have built
two powerful AIs in Acai.

Riky  6:18
One is an AI supervisor.

Riky  6:20
This AI supervisor manages, escalates to
manager, understands performances,

Riky  6:26
monitors everything that
is happening, as well as

Riky  6:29
route contact to the right
agent or to the right AI.

Riky  6:33
Instead, the other powerful AI is an AI
travel agent.

Riky  6:36
This AI travel agent is kind
of an AI version of a very

Riky  6:40
senior back office agent and
a very senior front office agent.

Riky  6:44
So what it can do is change, cancel,
do all the complex manual reissue,

Riky  6:48
manual refunds, as well as bookings or
going to the airline website, interacting

Riky  6:53
with them, with their website,
making check-ins, understanding

Riky  6:57
policies, anything that an
experienced agent does.

Riky  7:00
And the way you can use this is to either
help season agents to be faster so that

Riky  7:06
they can focus more on the empathetic
connection with the traveler or to empower

Riky  7:10
younger, non-GDS-trained agents to perform
their job, because we do all the complex

Riky  7:15
stuff interacting with the GDS and
airlines, hotels, as well as we have

Riky  7:20
created an agentic, MCP-based
version of it so that any

Riky  7:25
other software in the travel
industry can connect to it.

Riky  7:30
And so this is kind of the start of this
new travel operative system.

Riky  7:34
We're going to build more and more
agentic-based software according to what

Riky  7:37
our customers need and connect them all
together and taking also a very open

Riky  7:41
approach so that all the other third-party
provider and software in the industry can

Riky  7:48
connect to and together we can make music,
if that makes sense.

Riky  7:51
Why I decided to build again?

Elena  7:52
Yeah, it does.

Elena  7:53
You actually made it sound quite simple,
but we know how complex it can get when

Elena  7:57
you get into the nitty-gritty of the
travel operating system.

Elena  8:00
So now, Gabriel, from the other side of
the table, as someone working in a larger

Elena  8:04
scale operation like
Amadeus, where do you see the

Elena  8:06
friction when it comes to
implementing transformative AI?

Gabriel  8:09
I would say, more than frictions, I think
this is a real impact that we are seeing.

Gabriel  8:15
Something that is here to stay.

Gabriel  8:17
You can look at that, and this is the way,
the angle we take is from two different

Gabriel  8:21
dimensions, which is a little bit echoing
what Riky was saying.

Gabriel  8:25
Because we are looking at the efficiency
dimension, because we know that AI will

Gabriel  8:29
allow us to streamline various
processes, that it will allow us

Gabriel  8:32
to remove redundancy tasks,
bring automation to another level.

Gabriel  8:37
And we have, parallel to this,
the business dimension, where we know also

Gabriel  8:42
that new players such as PDAs and AI in
general might redistribute a little bit

Gabriel  8:48
who's going to be what on the entire value
chain.

Gabriel  8:50
And even potentially redistribute certain
entire value chain steps.

Gabriel  8:55
So AI is clearly something where we pay a
lot of attention.

Gabriel  8:59
And again, really from these two angles on
travel distribution, trying to see what

Gabriel  9:03
does that mean for our
day-to-day, and also what

Gabriel  9:06
does that mean from a
business point of view.

Gabriel  9:08
That's the angle we take.

Elena  9:09
All right.

Elena  9:10
Now, Riky, I'm always
excited about interviewing

Elena  9:12
serial entrepreneurs with previous exits.

Elena  9:15
And your background is quite impressive.

Elena  9:16
You were already mentioning about it.

Elena  9:18
But you sold your
previous company called

Elena  9:21
30 Seconds to Fly and became
head of AI at Amex GBT.

Elena  9:24
Can you tell us a bit more
about that initial company?

Elena  9:27
And also, why did you
choose to build again?

Elena  9:29
Maybe even what gap did
you see that was still unsolved, 

Elena  9:32
even after all that insider
experience that you had?

Riky  9:39
Well, you know, I wasn't thinking to go
back to travel.

Riky  9:42
You know, travel is an industry I deeply
love, but I also felt back then after

Riky  9:48
selling 30 Seconds to Fly that the
industry is very complex, very fragmented,

Riky  9:53
full of edge cases, and based on the
technology that we had available before,

Riky  9:59
it was difficult to deliver to all the
wishes that everybody had in the industry.

Riky  10:04
So I was thinking to do something else in
a different industry.

Riky  10:06
But then, as you said, in 2022, LLMs
came out, and that changed the whole game.

Riky  10:14
Finally, all our wildest dreams became
possible.

Riky  10:18
And I think I envisioned how this
technology would evolve and how this

Riky  10:24
technology would create an opportunity for
all this software to connect together.

Riky  10:28
And I thought this
was a once-in-a-lifetime

Riky  10:31
opportunity to go and help
the industry move forward.

Riky  10:35
So I decided to leverage all that
expertise and build a new company.

Riky  10:38
The vision I saw is that I think we have
built a lot of good, very good software in

Riky  10:44
the past 20, 30 years altogether,
which, anyways, is deterministic software.

Riky  10:49
So it's based on if, else, and loops.

Riky  10:52
So this works very well for a lot of happy

Riky  10:54
paths, but the industry
is full of edge cases.

Riky  10:56
And so a lot of time we
go into a stuck situation.

Riky  11:00
You cannot change your
whole customer service.

Riky  11:03
The symptom of this is that all this
deterministic software, especially when it

Riky  11:08
connects together through these structured
APIs, often has limitations.

Riky  11:13
Again, it's almost like you have very good
musicians.

Riky  11:17
When you try to make them
play music together, in the end,

Riky  11:21
they make a dissonant
music out of beat, out of tempo.

Riky  11:25
This is kind of where we are.

Riky  11:26
And it's not because
we have any fault or not.

Riky  11:29
It's just the limitation
of this deterministic

Riky  11:32
technology in an industry
that is very complex.

Riky  11:35
And I think, for the first
time, we have an opportunity

Riky  11:37
with agentic software
to revolutionize this.

Riky  11:42
So what I see in the industry is that
we're going to start having agentic-based

Riky  11:48
systems in a lot of the building blocks of
the travel tech industry, or at least a

Riky  11:54
hybrid approach, like augmenting existing
systems with some additional intelligence,

Riky  11:59
some additional cognitive power.

Riky  12:01
And when this system becomes smarter,
and we make them talk together like a

Riky  12:06
conversation between humans, similar like
if you put agents from different companies

Riky  12:11
and different parts of the organization
talking together, they will do the same.

Riky  12:14
They will talk with each other.

Riky  12:15
And then they will solve all
these complex situations that

Riky  12:19
would create a bug or an
issue or a dissonant experience.

Riky  12:23
And I think it's especially interesting,
not only when you have in the same entity,

Riky  12:29
like a TMC or OTA or an airline,
all this software talking together in a

Riky  12:35
smart, flexible way, but also when you
start having different entities talking

Riky  12:39
together in this smart, flexible
conversation.

Riky  12:43
So, for example, you have a corporate
having an agentic conversation with a TMC.

Riky  12:48
The TMC has an agentic conversation with
the GDS or with the hotels and airlines.

Riky  12:52
And anything that is very complex gets
solved.

Riky  12:56
Any complex change or cancellations or
other complex edge cases get solved.

Riky  13:02
So I think the ultimate beneficiary of
this is, first of all, the traveler,

Riky  13:06
because the traveler will finally have
this beautiful experience that they crave.

Riky  13:11
But also a lot of players in the industry,
including TMCs, OTAs, airlines and GDSs,

Riky  13:17
they will benefit from not
having to hire more and more

Riky  13:21
customer service rep
that is very difficult to find.

Riky  13:24
And also decreasing the cost in IT
manpower because all these smart systems

Riky  13:28
don't need as much maintenance as the
deterministic ones.

Elena  13:31
All right.

Elena  13:31
So, yeah, there was really a new
opportunity, right?

Elena  13:34
A new door, a new
opportunity for creating

Elena  13:37
something even bigger
than you created in the past.

Elena  13:40
But now, again, with
you, the question,

Elena  13:44
you walk away from a 
major role at Amex GBT,

Elena  13:48
Which was, by the
way, quite a bold move.

Elena  13:52
What was the tipping point for you?

Elena  13:53
Do you think it's just harder to build

Elena  13:55
meaningful innovation
within a big corporation?

Riky  13:58
Good question.

Riky  14:00
I think it was threefold.

Riky  14:02
Like on one end, yes, as you said,
I love building and in a corporation,

Riky  14:08
building is slower.

Riky  14:09
This is normal.

Riky  14:10
A corporation needs
to defend all the integrity

Riky  14:15
and the safety of a high
multiplicity of customers.

Riky  14:18
So, of course, it needs to be slow to make
sure this is protected.

Riky  14:24
A startup can bend rules, can risk more
because it has a smaller customer base.

Riky  14:32
As well, if the startup is smaller,
so all the brains...

Riky  14:37
not all the brains, like
everything is connected in

Riky  14:40
a lower amount of brains,
so everything gets faster.

Riky  14:42
So I would say in a
startup, building innovation

Riky  14:45
was at least 10x faster
than a corporation.

Riky  14:47
So, yeah, I love to build.

Riky  14:49
I love to build companies.

Riky  14:49
I love to build products.

Riky  14:51
And naturally, I go back to building in a
smaller organization.

Riky  14:57
But yeah, I think it was also threefold.

Riky  15:00
As I said before, the second
reason, I thought this was a

Riky  15:03
once-in-a-lifetime opportunity
to reinvent the industry.

Riky  15:07
And I think I had a lot of experience
there, so I wanted to contribute.

Riky  15:11
And third, I would say, I had a dream of a
new company culture, which I've been

Riky  15:19
cooking for the previous five to seven
years.

Riky  15:24
And this company culture was an integrated
approach between a capitalistic structure

Riky  15:31
that grows, as well as a
deep ethical framework

Riky  15:34
that allows for people to
become better individuals.

Riky  15:37
So this was inspired in a lot of practices
I went into between Zen, cognitive

Riky  15:45
behavioral theories, Tibetan Buddhism,
and other philosophies and religions.

Riky  15:50
I tried to find the
intersection between all

Riky  15:52
of them and try to
find the core of wisdom.

Riky  15:56
And so I wanted to build a company
spreading this message and creating this

Riky  16:00
integrated approach where the capitalistic
structure grows, the startup grows.

Riky  16:06
And this creates fition, so opportunities

Riky  16:09
to reflect, to create difficult situations.

Riky  16:13
Like, you know, you have new people joining
and people need to reinvent themselves.

Riky  16:15
These are all difficult situations.

Riky  16:17
And then the ethical approach makes them
reframe the situation as opportunity for

Riky  16:22
growth, both as a company, both as an
individual.

Riky  16:25
And the result of this is, in the end,
having, like, raising people that become

Riky  16:31
beautiful, empowered, inspiring leaders,
husbands, wives, and friends.

Riky  16:38
So the company, in the end, becomes a
playground for all these people to become

Riky  16:41
better individuals and
bring all this progress

Riky  16:44
forward to whatever
adventure they go to.

Riky  16:46
So this, you know, makes
me very happy every day.

Riky  16:48
I think it's a beautiful approach and
hopefully they will write books about it.

Riky  16:53
And so I wanted to build again and also to
make this beautiful experiment.

Elena  16:58
All right.

Elena  16:58
Yeah, I think I stuck with your idea of
having a small boat on and pointing out

Elena  17:03
the rowers, the few rowers
towards the same direction.

Elena  17:06
So that's when it
becomes easier to innovate.

Elena  17:10
Gabriel, back to you.

Elena  17:12
Do you agree with Riky on this?

Elena  17:14
What do you think are the challenges of
innovating within a big corporation?

Elena  17:17
All right.

Gabriel  17:18
Yeah, I agree in a sense also that the
referential at the beginning is not the same.

Gabriel  17:23
Because when you are a startup,
you start something.

Gabriel  17:27
When you are a big corporation,
you have something in your hands where you

Gabriel  17:30
need to combine the new stuff coming,
how you want to move, but also put that

Gabriel  17:36
into the context of the entire portfolio,
for example, that you have.

Gabriel  17:40
Or as you said, Riky, with the relationship
that you have already with customers.

Gabriel  17:44
So from that angle, the corporation is
starting from a different environment,

Gabriel  17:50
I would say.

Gabriel  17:50
And that creates mechanically also,
as you said, that in a startup,

Gabriel  17:55
very few people can be connected to deal
with a broad scope.

Gabriel  17:59
In a big corporation, you
need to have somehow a certain

Gabriel  18:03
inertia because of the size
of what you are dealing with.

Gabriel  18:07
So somehow there is a fundamental
difference here.

Gabriel  18:09
And somehow you can see that talking about
AI happening in real life also with PDAs.

Gabriel  18:14
OpenAI starts from scratch.

Gabriel  18:16
So they are moving at the
pace that they consider they

Gabriel  18:19
can move, also releasing
things relatively rapidly.

Gabriel  18:24
When Google does not, with
Gemini, because Google, Gemini

Gabriel  18:28
is a question also in relation
to their existing business.

Gabriel  18:32
So whatever move they do, they need to
project that into a different context and

Gabriel  18:37
to make the movement
that makes sense, relative

Gabriel  18:40
also to all what they
have to deal with.

Gabriel  18:42
So this is why there are two very
different dynamics.

Gabriel  18:46
So in corporations, still though,
when we talk about AI, what I have

Gabriel  18:49
learned, because I
was like Riky when my

Gabriel  18:52
daughter showed me
ChatGPT, it was like in 2022.

Gabriel  18:56
When I saw that, I had a reaction where
after the second question that I asked,

Gabriel  19:00
which was something like, give me the NDC
strategy that GDSs should follow.

Gabriel  19:05
Just looking at how poor is going to be
that answer.

Gabriel  19:09
And I had the answer and
then I thought to myself,

Gabriel  19:12
damn God, iRobot and
Terminator are arrived.

Gabriel  19:15
Okay, more or less.

Gabriel  19:16
And I felt like, ooh,
this is a new animal.

Gabriel  19:18
Something is cooking here.

Gabriel  19:19
So I created immediately an incubation
team on my side, just at the beginning to

Gabriel  19:24
try to understand what was coming,
because that was not obvious.

Gabriel  19:27
This is a different thing compared to what
we have seen so far.

Gabriel  19:30
And it comes with a very big dimension of
speed that we have never seen before.

Gabriel  19:36
Okay.

Gabriel  19:36
To that point, in terms of evolution,
is it going to reach a peak?

Gabriel  19:40
A lot of people say yes, that now we are
at a moment where because of cost

Gabriel  19:45
constraints, because of energy
constraints, because of certain limits

Gabriel  19:49
that you can have in the
LLM and RAG model, certain

Gabriel  19:53
people start to say, well,
we're going to see a slowdown.

Gabriel  19:55
You have other people saying the opposite.

Gabriel  19:57
The reading is extremely difficult.

Gabriel  19:58
So this is why for corporations,
this is an interesting animal to play

Gabriel  20:02
with, because you need to
react rapidly to that animal.

Gabriel  20:06
But be careful, you can't get what
you want because you can react rapidly.

Gabriel  20:11
However, your reaction might
not be the right one, because

Gabriel  20:14
it's a very difficult reading
that you have in front of you.

Gabriel  20:17
And you might react strongly
on something where you learned

Gabriel  20:21
after with the hard life that
it was not the right move.

Gabriel  20:24
So it's important also not to panic.

Gabriel  20:27
Because we see a lot of things that are
communicated, it's difficult to read.

Gabriel  20:32
So you need to pay
attention to the signals, stay

Gabriel  20:35
highly connected to
what's going on, for sure.

Gabriel  20:38
One reaction that you must have is a must
have is what is available, use it.

Gabriel  20:43
Otherwise, you're going to be behind.

Gabriel  20:45
Because this is so powerful
that not embracing that

Gabriel  20:49
technology will become
very rapidly a handicap.

Gabriel  20:51
I had someone over the weekend,
I was listening to something on YouTube,

Gabriel  20:55
where someone said, the problem is not
that people will lose their job.

Gabriel  20:59
Because to lose your job, it means that
the AI must be 100% human capability.

Gabriel  21:04
Debatable that this is an easy one to do.

Gabriel  21:07
Now, AI is starting to do
certain things that allow you

Gabriel  21:10
to be quicker, allow you to be
almost an augmented person.

Gabriel  21:15
So if you don't embrace that
technology, the job that are

Gabriel  21:18
at risk are the job of people
not embracing that technology.

Gabriel  21:22
To me, this is more what
is likely to happen when

Gabriel  21:26
people get fear about
loss of job, blah, blah, blah.

Gabriel  21:28
Because, well, no, if you use that
technology, if you know how to use it,

Gabriel  21:33
in any case, you will be part of the
marvels, having the augmented power.

Gabriel  21:37
Okay, and this is to me what is at stake.

Gabriel  21:39
So long story short, different dynamic for
corporations compared to startups,

Gabriel  21:45
a different angle to look at it.

Gabriel  21:47
But with the same
sense of urgency, okay?

Gabriel  21:50
With maybe a reaction that will be
slightly different because you are not

Gabriel  21:53
navigating in the same context and
ecosystem.

Riky  21:57
If I may, Gabriel, I think I really agree
with you on the super-powered employee.

Riky  22:04
I think what we're seeing in many
industries is the rise of this full-stack

Riky  22:09
employee, this employee that can do so
much more.

Riky  22:12
Like, it's not anymore vertical,
but it's horizontal.

Riky  22:15
This is empowered by AI to do all the
vertical tasks.

Riky  22:18
And I think also in travel, we will land
into a full-stack agent.

Riky  22:23
You know, you don't have anymore the
concierge, the front office agent,

Riky  22:26
the back office agent, et cetera,
et cetera.

Riky  22:28
You'll have one empathetic individual that
can offload all these tasks to other AIs

Riky  22:34
and can focus more
on providing a fantastic

Riky  22:38
connection service to the
traveler they're talking to.

Gabriel  22:41
AI will bring, finally, the emphasis on
how huge is the value of what human can do.

Gabriel  22:50
This is what is at stake at the end of the
day.

Gabriel  22:53
The value of all what you mentioned about
type of services, human dimension of

Gabriel  22:59
things, this is going to raise a lot in
terms of value, for sure.

Riky  23:05
Yeah, and it's going to be fun.

Riky  23:07
I think it's going to be a very fun world
working.

Gabriel  23:09
It's going to be fun.

Elena  23:11
Going back to you, on hearing your
story about meeting I, because most of

Elena  23:15
the times our ventures team scouts the
companies we're investing, but we

Elena  23:19
definitely need sponsorship and alignment
with the corresponding business unit.

Elena  23:23
But in this case, when you first
encountered I, what stood out to you?

Gabriel  23:27
Well, first, we are connected.

Gabriel  23:30
Look at just the conversation we have.

Gabriel  23:32
Everything that Riky is saying,
I think the same.

Gabriel  23:35
So, somehow, this is also very important
in a conversation like that, that is

Gabriel  23:40
extremely versatile depending on the
people you have in front of you,

Gabriel  23:43
the level of knowledge.

Gabriel  23:45
There is a little bit of crystal ball in
that conversation, also, let's be honest.

Gabriel  23:51
So, talking to Riky and looking at what
Acai is doing, we have a similar crystal

Gabriel  23:57
ball, somehow, even
though after the road might

Gabriel  23:59
be different on the way
we want to tackle that.

Gabriel  24:02
But this is also about complementarity.

Gabriel  24:04
And this is what I see very much.

Gabriel  24:06
We just discussed about what a corporation
has to face in terms of inertia,

Gabriel  24:12
in terms of way of behaving, okay?

Gabriel  24:14
And we see what Riky
was saying on his side.

Gabriel  24:18
So, there is a very, very interesting...

Gabriel  24:21
We have very interesting ingredients to
mix, okay, in that conversation.

Gabriel  24:25
As soon as you are connected in the same
way, and this is the case.

Gabriel  24:29
We look at things from the same angle,
and this is very important.

Gabriel  24:34
Plus, I would just add on top of that,
well, you heard Riky, you have a

Gabriel  24:40
fantastic background in the travel
industry.

Gabriel  24:42
And travel is not an easy subject.

Gabriel  24:45
Otherwise, PDAs would have cracked that
already.

Gabriel  24:47
And we see that PDAs
move in travel so far, but

Gabriel  24:51
they move essentially on
efficiency conversations.

Gabriel  24:55
How can you optimize the way you work,
blah, blah, blah, and so on.

Gabriel  24:59
But none of them yet has shown a very big
movement on the travel business of things,

Gabriel  25:04
because this is a very complicated world.

Gabriel  25:07
And this is where Riky's background can
really help in moving rapidly,

Gabriel  25:13
because we talk that language,
the two of us.

Elena  25:15
Exactly.

Riky  25:16
The love is mutual, guys.

Riky  25:19
Love all you guys.

Riky  25:20
I think Amadeus are very, very smart
people.

Riky  25:23
It's a big organization, but I think it
can still be agile.

Riky  25:26
All the interactions we had
were marvelous and a lot of

Riky  25:29
respect for everybody there
and everything you're doing.

Elena  25:32
I'm glad to hear that there is alignment.

Elena  25:34
That's exactly what happened when we
decided to invest in you guys.

Elena  25:39
So, yeah, I agree that the most important
characteristic for early stage investors

Elena  25:43
is, and that I would say everybody should
look into, is to invest into strong teams, 

Elena  25:48
really able to be resilient
no matter what happens, 

Elena  25:50
and teams that are
able to pivot if needed.

Elena  25:54
So now talking about that investment
angle, Riky, you met Suzanna,

Elena  25:58
our head of Amadeus Ventures, many years
ago.

Elena  26:00
Back then you had already raised capital
from a strategic, so basically corporates

Elena  26:04
from the travel industry like FCM or
JetBlue Ventures.

Elena  26:07
So once again, very experienced in dealing
with corporate venture capitals.

Elena  26:10
And this time you got to raise capital
from us, from Amadeus.

Elena  26:13
What do you think are
some of the challenges

Elena  26:15
of raising from a
corporate venture capital?

Elena  26:17
And what advice would you give to founders
entering this kind of collaboration?

Riky  26:20
Very interesting question.

Riky  26:23
By the way, Suzanna, she's great.

Riky  26:24
She's a fantastic corporate VC.

Riky  26:27
She's fun.

Riky  26:28
She's a great negotiator.

Riky  26:30
She knows the industry well.

Riky  26:31
She knows Amadeus well.

Riky  26:32
It's great to work with her.

Riky  26:34
Just, you know, you were
saying about, I received

Riky  26:36
the investment from
FCM JetBlue, also Elal.

Riky  26:38
This was in the previous company.

Riky  26:39
But they were corporate VCs, as Amadeus is
a corporate VC.

Riky  26:44
And I think if we need
to generalize, corporate

Riky  26:48
VCs, generally, they look for information.

Riky  26:52
They look for research.

Riky  26:53
They look for products that can complement
their offering.

Riky  26:58
They look for an M&A pipeline.

Riky  27:01
And they care generally less of the pure
financial return.

Riky  27:06
While instead, the institutional VCs,
they only look for the financial return.

Riky  27:11
And on the founder's side, you know,
sometimes you're more in a situation to

Riky  27:14
side with the objective of
the corporate VC, sometimes

Riky  27:16
more on the objective
of the institutional VCs.

Riky  27:19
So there is not a right or wrong.

Riky  27:21
You know, in business, everything is very
complex.

Riky  27:24
And I think for us, it was very important
to make this partnership.

Riky  27:29
We believe, again, as was saying,
that we're very aligned.

Riky  27:33
Hopefully, it's going to be a nice
symbiosis in which we can experiment,

Riky  27:38
fail fast, and bring back some of the
learnings back to Amadeus.

Riky  27:43
And at the same time, we're
looking forward to leverage all

Riky  27:46
the network and experience
connectivity that Amadeus has.

Riky  27:51
And then, you know, you were saying
challenges, challenges to partner with a CVCs.

Riky  27:58
I think, you know, like, it's
normal that Amadeus needs

Riky  28:01
to push for getting as much
information as possible.

Riky  28:04
And at the same time, the startup needs to
be as independent as possible.

Riky  28:09
And then, you know, you land into,
like, a healthy middle ground in here.

Riky  28:15
And I think that's why it also took,
you know, one year to negotiate this deal.

Riky  28:19
It was a long path.

Riky  28:22
And congrats again, everybody,
for all the hard work there.

Riky  28:26
Yeah, so if we need to give advice to
anyone else partnering with corporate VCs,

Riky  28:31
I would say, you know,
take your time, because

Riky  28:34
there is a healthy middle
ground for everybody.

Riky  28:37
Don't rush into anything.

Riky  28:38
And in the end, land into this perfect
deal that can protect everybody,

Riky  28:43
protect the independence of the startup,
any possibility for competition or not.

Riky  28:50
And in the end, you forecast all these
scenarios and create a good, healthy deal.

Elena  28:55
And Gabriel, from your seat,
jumping into more of the partnership

Elena  28:58
side of things, because this
goes beyond just investment.

Elena  29:01
What was it about I that aligned with
Amadeus operational strategy?

Elena  29:04
And how do you envision
integrating their technology

Elena  29:07
into the broader
ecosystem within Amadeus?

Gabriel  29:10
Well, when we first met
with Riky, we started

Gabriel  29:13
to share with each
other what we are doing.

Gabriel  29:16
We've seen that on the Acai side,
they have moved on the two dimensions that

Gabriel  29:21
I was mentioning earlier, with proof of
point that they have been successful on

Gabriel  29:26
the efficiency angle, how you can automate
more, how you can use AI also to replace

Gabriel  29:31
you on certain things that are very low
value, or that can handle after the

Gabriel  29:36
relationship with
customers, more from a call

Gabriel  29:38
center point of view, but
also on the business side.

Gabriel  29:42
So what we agreed immediately was to
explore potential use cases, because this

Gabriel  29:48
is something that we have done on our
side, of course, is to identify the use

Gabriel  29:51
cases that we consider with high value,
and to ensure that we can benefit from the

Gabriel  29:57
collaboration we were somehow talking
about earlier, and that Riky just also

Gabriel  30:00
mentioned, to test rapidly
certain things, identify

Gabriel  30:03
together the bets, because
we talk travel, both of us.

Gabriel  30:07
So what are the big bets, in fact,
that we should focus on?

Gabriel  30:10
How we can rapidly probably experience the
value we can get out of these bets,

Gabriel  30:14
if they pay back or if they make sense.

Gabriel  30:16
And yes, this is to me, the big synergy
that I see is that we can work together on

Gabriel  30:23
proof points on AI use cases that will
bring us to high value solution before

Gabriel  30:29
moving to industrialization, because
Amadeus is giving that also is the

Gabriel  30:33
possibility to move at large scale in
terms of industrialization.

Gabriel  30:36
But you better have a certain level of
confidence when you press the button,

Gabriel  30:40
when you're at that stage, because the
consequences are important in terms of

Gabriel  30:44
investment, in terms of movement,
momentum that you need to give,

Gabriel  30:48
because you talk about many markets.

Gabriel  30:50
And this is where we decide that's going
to be great to work together, because we

Gabriel  30:54
have that complementarity
that I was mentioning

Gabriel  30:55
earlier, and we are already
going to explore use cases.

Elena  30:59
All right, so now this is my favorite part
of the session, when we talk about the

Elena  31:04
vision, not only where all our efforts are
heading to, but also in a broader sense, 

Elena  31:09
where do you think, where do we think AI
will take us?

Elena  31:13
So let's start with you, Riky.

Elena  31:14
How do you think I will roll out?

Elena  31:17
And what will it be in 5 to 10 years from
now?

Riky  31:20
Yeah, I think it's similar to what I was
unpacking before.

Riky  31:23
I see an opportunity where a lot of the
building blocks of the travel tech

Riky  31:31
infrastructure will
start to be rewritten in

Riky  31:34
a genetic mode, or it
can be a hybrid of that.

Riky  31:39
And in the end, all this software will
connect together in this smart,

Riky  31:42
flexible, dynamic composition way with
this reasoning power, same as humans.

Riky  31:47
And in the end, we create
this harmonious travel

Riky  31:51
tech infrastructure
where everything works.

Riky  31:55
And the result is that travel is happy,
everything works, and we're going to also

Riky  31:58
decrease operating costs, decrease all the
nasty work to manage a software that is

Riky  32:03
too complex, deploying a software that is
too complex, or managing difficult

Riky  32:08
scenarios in travel where we don't have
agents to solve those issues.

Riky  32:13
But it's interesting to say that not all
parts of the travel tech infrastructure

Riky  32:20
are actually ripe for disruption with
agentic technology.

Riky  32:25
I would say the advantages of agentic
technology is that there is cognitive power.

Riky  32:32
The disadvantages is that it is way more
costly to operate.

Riky  32:38
Rarely it can make mistakes according to
how good you are to use it.

Riky  32:42
And it can be slower.

Riky  32:44
There is a spectrum of technology where
agentic can be applicable.

Riky  32:50
If we make an example on, let's say,
back-office finance systems, these systems

Riky  32:57
need to be extremely accurate and
extremely fast.

Riky  33:01
So I would say deterministic technology
here will withdraw.

Riky  33:05
If I take instead maybe a
mid-office, here you still need

Riky  33:10
something very accurate and
very fast and cost-effective.

Riky  33:13
So I would say deterministic technology
will be there.

Riky  33:15
But for anything that is edge cases or
errors or things that fall out of cues,

Riky  33:22
then you will have a backup agentic
mid-office that will fix those situations.

Riky  33:27
And for anything instead that is extremely
complex, very fragmented, lots of edge

Riky  33:33
cases, here is where agentic technology
gives the best.

Riky  33:37
A manual change or cancellation,
this is very costly to operate.

Riky  33:42
So here agentic is perfect.

Riky  33:45
Interestingly enough, there are a lot of
startups that do the agentic booking.

Riky  33:52
They focus on the booking side.

Riky  33:55
And this is a tricky one
because, on one hand, yes,

Riky  34:00
agentic can help making a
better booking experience.

Riky  34:04
But at the same time, the revenues on a
booking, let's say in corporate travel,

Riky  34:08
is what?

Riky  34:09
$4 or $5 is what they bill for an
automated booking on an OBT.

Riky  34:13
And if you think agentic technology doing
bookings, at the current costs,

Riky  34:19
you would have an electiity cost, an

Riky  34:22
inference cost of $5,
$10, $20 for each booking.

Riky  34:25
So who's going to pay for this?

Riky  34:27
So this is kind of maybe a warning to all
the startups out there that do bookings.

Riky  34:33
Be very careful of your unit economics
because it might not work.

Riky  34:38
Check what is all the distribution,
all the pipeline, and how the money is

Riky  34:42
flowing to the industry
to understand if the

Riky  34:44
agentic booking actually
can be scalable or not.

Riky  34:47
So again, not everything is ripe for
disruption with agentic, but the more the

Riky  34:52
inference costs will go down, the more
data centers we will build, the more these

Riky  34:57
models will be cached and fast,
the more things we can build.

Riky  35:02
And it's going to be a very interesting,
fun future.

Elena  35:05
Yeah, I agree.

Elena  35:06
Very excited.

Elena  35:06
Let's see what happens.

Elena  35:07
Back to you, Gabriel,
with just that big vision

Elena  35:10
perspective on what
we were talking about.

Elena  35:12
We are already seeing the
impact of AI across industries, 

Elena  35:15
but do you envision a
bigger disruption due to AI?

Gabriel  35:18
I would say the very first one to me has
been really the arrival of the PDAs,

Gabriel  35:22
like ChatGPT, Gemini, that I was
mentioning, or perplexity.

Gabriel  35:26
These animals bring already
a significant disruption.

Gabriel  35:29
And the way I like to
look at it is AI will bring

Gabriel  35:34
certain evolution, certain
disruption, and revolution.

Gabriel  35:40
We try to look at it from that angle.

Gabriel  35:42
So, for example, talking about agentic,
and I think it's important to keep in mind

Gabriel  35:50
what Riky was saying,
which is it has an impact

Gabriel  35:54
on speed sometimes, it
has an impact on conception.

Gabriel  36:00
But all in all, I would say
agentic brings the orchestration

Gabriel  36:05
capability to a level that
is still to be evaluated.

Gabriel  36:12
But depending on how much you can
orchestrate, because if you think,

Gabriel  36:16
if you push the extreme
on orchestration, our

Gabriel  36:19
brain, ourselves, is
orchestrating permanently.

Gabriel  36:21
This is what we are
doing about many things.

Gabriel  36:23
So, agentic is as a
connective dimension, meaning

Gabriel  36:27
that you never know
how far that can evolve.

Gabriel  36:31
That's to me a little bit of
a question, where agentic

Gabriel  36:34
can move from one of the
three can be an evolution.

Gabriel  36:37
I would bet more on at least a disruption.

Gabriel  36:40
But can it go down to a revolution,
depending on how far we can go with that.

Gabriel  36:44
But to me, the biggest
paradigm shift that is to come

Gabriel  36:49
in front of us is the arrival
of robots in the game.

Gabriel  36:54
Not robots as such, because we talk about
robots that can do certain things for us,

Gabriel  36:59
blah, blah, blah.

Gabriel  37:00
No, robots in a sense
that today the feed

Gabriel  37:04
received by the AI in
general comes from humans.

Gabriel  37:09
You are chatting, what you say,
the AI learns from it.

Gabriel  37:12
We have created documents, the AI learns
from it.

Gabriel  37:17
So, we are the input and the eyes and the
understanding of the world for the AI.

Gabriel  37:23
As soon as the robots
arrive, even if it's a robot that is

Gabriel  37:26
supposed just to take care
of your grass in your garden.

Gabriel  37:30
But something that we
start to give data to the

Gabriel  37:35
engine, to the animal, that
are not a feed from human.

Gabriel  37:39
That is a feed from certain things that
the AI will collect elsewhere.

Gabriel  37:44
And the AI then will have the possibility
to start to grow with even data to try to

Gabriel  37:51
start to understand what is the external
world.

Gabriel  37:55
Where today, the AI does not understand
what is the external world.

Gabriel  37:59
To me, if you ask me,
this is one of the biggest

Gabriel  38:02
paradigm shifts that can
happen again, another layer.

Gabriel  38:06
We had the PDAs that were kind of being
intelligent to answer our questions.

Gabriel  38:15
We have the agentic
coming now, allowing to do

Gabriel  38:19
actions and to orchestrate
to a level to be defined.

Gabriel  38:25
And as soon as you start to inject data
that is not coming from human.

Gabriel  38:29
That is coming from
external sources that the

Gabriel  38:33
AI will get in order to
grow its intelligence.

Gabriel  38:36
And because an AI wants permanently to
learn, that one to me is an interesting

Gabriel  38:40
one that will have to come in the picture
at one stage.

Elena  38:43
All right, so thank you both.

Elena  38:45
That was a very, very interesting
conversation.

Elena  38:47
That's a wrap for today's episode.

Elena  38:49
Hope you all enjoyed and got
inspired by this conversation.

Elena  38:52
Riky, Gabriel, thank
you both for your insights.

Elena  38:54
Now we know, one, AI is no longer
theoretical in travel.

Elena  38:57
It's operating and it's at the
core of the travel industry.

Elena  39:00
And two, this is just getting started.

Elena  39:02
Startups, the opportunity is here for you.

Elena  39:04
Corporations, need speed to market.

Elena  39:06
And Corps, this is your time to
be open-minded and collaborate

Elena  39:08
with the quick ones to
make this revolution happen.

Elena  39:11
To our listeners, thank you for listening.

Elena  39:13
And don't forget to subscribe
to Amadeus Travel Tech Talk

Elena  39:16
on Spotify or Apple Podcasts.

Elena  39:18
And visit amadeus.com/transformingtravel
to learn more about how

Elena  39:21
Amadeus is partnering to
shape what's next in travel.

Elena  39:24
Thank you all and see you soon.

Riky  39:26
Thank you, Elena.

Gabriel  39:27
Thank you Elena.
Bye-bye.

Transcribed by https://otter.ai