Travel Buddy with Switchfly

AI agents can already book travel directly. That raises a hard question for loyalty programs. If a customer books through an agent instead of a portal, does their loyalty status even factor in? Ian Andersen joins to unpack how agentic booking works today and how it optimizes almost entirely on price. He walks through the standards emerging to change that, including Google's Universal Commerce Protocol and Talon's Universal Incentives Protocol. Both are starting to make loyalty data machine-readable. The conversation covers what program and product leaders should prioritize now and why waiting has a real cost.

Connect with Switchfly
Website: https://www.switchfly.com/
LinkedIn: https://www.linkedin.com/company/switchfly/
X: https://twitter.com/switchfly
YouTube: https://www.youtube.com/@SwitchflyOfficial

Chapters
[00:00] Introduction to agentic AI booking and why loyalty programs are affected
[00:01] Agentic booking goes mainstream: Google, Sabre, PayPal, and Mindtrip
[00:04] The optimization blind spot: price and availability, but not loyalty
[00:10] Emerging standards: Google's Universal Commerce Protocol and Talon's Universal Incentives Protocol
[00:20] The API-readiness payoff: build once, plug into whichever standard wins
[00:22] AI readability as the next answer engine optimization frontier
[00:25] Context loss from travel agent to digital platform to AI agent
[00:31] Quantifying value beyond price, and the two-user problem
[00:41] Closing takeaway: even ignoring this is a plan, just an expensive one

Creators and Guests

Host
Ian Andersen
Demand Gen & Marketing Operations Manager

What is Travel Buddy with Switchfly?

See more at Switchfly.com

Welcome to Travel Buddy,
presented by Switchfly.

In this podcast, we talk about all
things travel, rewards, and loyalty.

Let's get to it.

Brandon Giella: Okay, welcome
back to Travel Buddy by Switchfly.

We are gonna talk today about how your
loyalty program is invisible to AI.

As we know, AI is changing everything.

It is the topic that virtually every
podcast on Earth is about at the

moment right now, which is fine.

But, uh, there is a, a change in the way
that people are booking travel, and we

wanted to talk about it, so we've got
Ian here, uh, to, to give us the lowdown.

And so essentially what I'm understanding
is in Google AI, Google's, you know,

AI Overviews feature, you can now
book a hotel within that feature.

Which raises the question, what if people
don't actually go to your loyalty program

portal or your travel booking portal
or the actual thing where we're going

and choosing the hotel and clicking on
dates and making the selection ourselves?

What if it's becoming a lot more
What if we're just talking to some of

these tools and it's booking for us?

I know there are some tools out there, I
think Muse of them that I've been reading

about, and then there's another one that
I just read about, where you literally

are, like, texting or talking to your AI
agent, and it's going and booking things

for you using plumbing of other tools.

Who knows how this is happening?

We don't actually know.

So Ian, you know.

Talk to me

Ian Andersen: That's

Brandon Giella: is going on.

What is the shift that you're seeing here
and, and what is this thinking about?

H-how do you guys think about travel
now in the age of AI at Switchfly?

Ian Andersen: Sure.

Um, well, thanks, Brandon.

Uh, and thanks for the confidence
in, in believing that I know 'cause,

Brandon Giella: everything, Ian

Ian Andersen: if the, uh, if the engineers
at, uh, Anthropic and OpenAI and them

don't know, then I'm for sure lost.

But, uh, no, I, I mean, you, you,
you hit the nail on the head, right?

Is, is…

And this has obviously been, on the
roadmap and, and in people's sort of

like long-term horizons for what AI would
get to as far as not only, providing

information, but actually acting upon
that information on, on someone's behalf.

you know, we've, we've all known
this kind of thing is coming.

I think it's coming faster than
people might have necessarily thought.

I mean, just looking back at the
conversations we've had over the

past couple years on AI, you know,
it's, uh, the, the progress has

been, uh, astounding, you know, for
how rapidly things have changed.

and this is kind of the next step in it.

Like, like you mentioned, Google, as
of now, as of the end of August, an

agent can actually book on your behalf,
through the, the Google travel platform.

there's still some kinks to be worked out.

the, like you can't, it won't
book independent hotels, and

some other, other qualifiers.

But, you know, Sabre, PayPal, and
Mindtrip earlier this year would,

announce they have an end-to-end
agentic booking, pipeline, right?

that they have with their, uh, with
a lot of their internal connectors.

so this is clearly where things
are headed directionally.

So it, causes some problems for,
loyalty programs, as you mentioned,

like if they're booking direct,
if the agent's booking direct,

are they, you know, w-where does
the, your loyalty, program fit in?

Where, can a customer find
and use their rewards?

Things like that are a little
up in the air at the moment.

Right now, the way agentic booking is
working is it's optimizing based on, very

narrow restrictions on price availability.

You know, if you put in certain dates
and times and whatever your criteria

on, it's then essentially optimizing
for price, maybe one or two other

things like, number of layovers you
have, an air or, or, hotel, star rating

if, if you limit the, you know, I want
three-star and above or, or whatever.

but that-- those very narrow optimizations
leave out things like rewards points,

you know, booking preferences, historical
booking preferences, things like that.

and I think that's where this next step
in the, the evolution of AI booking

is headed, is being able to, uh, make
loyalty programs readable to, AI agents

and not only readable, but, but sort
of interchangeable, if you will, um,

between various AI platforms, and
creating some sort of standardization

that the AI will take into account
the rewards and loyalty aspect of the

booking they're trying to make, right?

And it becomes another Another variable
in the, the optimization, program, right?

So if, for example, the lowest cost
airline A is the lowest clo- cost, you

know, on the surface, but you have a
bunch of rewards, uh, you know, a bunch

of points saved up with airline B, which
then would make, if you factor that

into the equation, would then make the
airline B's, flight, more cost-effective.

how do we get-- how do we bubble up
that information to where the AI, is,

is making that determination as well?

and so, so program managers and, product
managers at travel, platforms that,

uh, when they're thinking about how
their system is set up, need to start

taking that into account and, uh, and
optimizing their platform for not only,

somebody going to book manually, but
also so an agent, reads their information

correctly and can make those, those
same evaluations when booking, as well.

Brandon Giella: It sounds simple to
say, you know, kind of in an idea of

like, uh, yeah, we need to optimize
our platforms and, and this is the,

the way that the technology's going.

But I have, uh, Switchfly CEO Noel
Outlaw in my head saying, "It's--

this is really complicated stuff to

Ian Andersen: Sure

Brandon Giella: It's kind of like seeing,
um, uh, you know, like you have an

iPhone and it's like, oh, I wanna like
something to my screen, or like I'm taking

a call here, sending a message here.

It's like wiring to my AirPods
or to my watch or what.

Like, all that is actually really hard
to do, and it's hard for a company

to map all those things together.

I guess, um, I, I guess the question
is like, is this like who sets

the, the standard or the plumbing?

Like h-h-how does this actually happen?

Like, 'cause I know like the
Switchfly platform, you guys have been

building this thing for twenty years

Ian Andersen: Yeah

Brandon Giella: all these, the,
the pieces together, and it's

really, really hard to get right.

What, h-how do you think about the
plumbing, but also like do you have

encouragement for people that might
be overwhelmed at the pace and the

speed of technology at the moment?

Ian Andersen: I don't know
if I have encouragement.

I have sympathy for people overwhelmed.

Brandon Giella: Sympathy

Ian Andersen: yeah, 'cause, 'cause
believe me, I feel it as well.

Brandon Giella: Yeah, I'm sure

Ian Andersen: the The heart of
the question, you're right, this

is incredibly complex stuff,
right, And it's, it's why, the…

It's why the AI hasn't
taken over yet, right?

Brandon Giella: Yeah,

Ian Andersen: of, uh, if, if it were
more simple, you and I would be out

painting in a field and enjoying
life while all our agents run around

taking care of everything else, right?

Already.

So,

Brandon Giella: I

Ian Andersen: so for, as simple as it
sounds, there are a lot of nuts and bolts

and, and like you mentioned, you know,
we've been in business for over 20 years.

you know, other booking platforms
have been out there for a long time.

Obviously, individual, hotel chains and,
and airlines and, uh, activity, you know,

amusement parks and things like that,
they've all been around a very long time

and, uh, over time have had to change
the way their models operate to sort

of fit in with whatever the current,
platform or situation technology, is.

And, and this is no different in that
it virtually always happens somewhat

organically in that somebody-- there'll
be two different solutions that come

up, uh, or, or avenues to do this.

one may or may not be slightly
better, you know, what- for whatever

the reason it tends to edge out the
other, and then a standard is created.

and then over time, everybody sort
of go- goes onto that standard.

I think the, the way to think
about it, uh, just for context,

Google has something called the
Universal Commerce Protocol, right?

It's their…

It's basically their guidelines
for, for anyone selling anything, on

the internet, through their website
or a platform or, or whatever.

If you set things up in this
particular format, it makes it easy

for Google to read and therefore,
easy for transactions to be made.

they are, they have something called
a identity linking capability in that,

a, they're trying to optimize for AI,
to be able to function with the, those

Universal Commerce Protocol, uh, protocols
that they've outlined, and make it

sort of a standard, across the board.

Google is Google, right?

They are the, still the sort of number
one search, uh, and, and entry point

for most people on the internet.

however, there's also
another one called Talon.one,

uh, another company called Talon.one,

that has launched a universal
incentives protocol, which is sort of

similar to the, the Google's universal
commerce, but it's kind of more d-

uh, refined towards loyalty programs.

the point I'm trying to make
is there are different avenues.

However, most of those things have a lot
in common as far as even if you go down

one route right now, the, the legwork
you're putting in to, to set up to go

down that route will be transferable
to a different route, you know.

if your-- Say, take your, your conversion
ratio, uh, the way you calculate

your, your currency, between, you
know, dollars and points or whatever.

if you're, if you're structuring that
in such a way to be readable by one

of these, these agents, by definition,
it's gonna be structured in a way for

something else to read it as well, right?

So, so say there, there is some startup
comes up with this magic solution

next year that all e-com, becomes
the standard for all e-com on the

internet, or the government regulates
a certain way to do it or, or whatever.

There are certain basic things that you
have to do no matter what, that will

be able to sort of plug into whatever
the plumbing, as you call it, to make

it readable for, for agentic commerce.

I hope that was

Brandon Giella: Yeah, I guess if I
could m- kind of back up, and maybe

there's some listeners that are, you
know, they're, they're loyalty leaders,

but they're not on the tech side, not
on the engineering or product side.

It's kind of like there's all these
kinda standard formats we have on the

internet that we don't think about
much anymore, like HTTPS, you know,

or, um, like a f- image format, there's
JPEG and PNG, or you could create a PDF.

But long ago, there were 100
different formats on all these things.

We've all, as a market, just sort
of solidified on one or two, and

now, like USB-C charging is now
like kinda the standard format for

Ian Andersen: Yep.

Brandon Giella: So we all
kinda onto one standard.

So you're saying the standard now
get a loyalty program, hotel booking

program, those kind of things into
at least one of these AI platforms,

there's a protocol that Google has
called Universal Commerce Protocol,

Ian Andersen: Mm-hmm.

Brandon Giella: and
then there's a Talon.one

that's like a more
specific version of that.

But these protocols are kinda getting
worked out now, and you can build toward

them, and that will help your loyalty
program to be readable to the AI.

Is

that--

Ian Andersen: yeah.

Brandon Giella: Is that right?

Ian Andersen: No, that's exactly right.

Um, there are far more protocols than
the couple I mentioned, but those are,

those are definitely industry leaders.

and they're-- it's a little, uh…

Actually, I think a long time ago, you
brought up this analogy that I think of

often about how, and I don't even remember
what the conversation was, but how, you

know, because the Romans built their
roads in such a fashion, the space shuttle

now has to be designed a certain way.

and it had to do with, you know, the
Romans built the roads, and that just

sort of became the standard for how
other European countries and eventually

United States built our roads.

And then because of that, you know, the,
the, because the space shuttle has to

be transported along the road system, by
the, by however method the Romans used to

determine their road size now determines
the design of the space shuttle.

And I always, I've, I think of
that analogy at random times,

because we, you know, we have to
live in the world provided to us.

And, the current state of things is
Google is, is king of the internet, right?

Like, I think AI over the last couple
years has, has taken some of that

away from Google as far as being,
kind of the search platform of choice.

But, but we still live in a
reality where the vast majority of

internet standards over the past
two decades have been determined

by what's best for Google, right?

and you and I work in marketing.

SEO is a, a very common term that
actually predates Google or definitely

predates when Google was the platform.

and it is, I mean, just
language agnostic, right?

It's search engine optimization.

But now it only means how to
optimize for Google, right?

Like, nobody really considers what other
search engines, and there are plenty

of other search engines out there.

It's almost not a non-factor of, of
any other optimization other than what

Google says is best for, for their,
for ranking on their algorithms.

So-

Brandon Giella: 98-- 95 or 98% of search

Ian Andersen: Yeah

Brandon Giella: You have like 1% on Bing
and 1% on DuckDuckGo and whatever, but

Ian Andersen: Yeah.

Brandon Giella: doesn't even matter.

Who

Ian Andersen: And whatever the,
what's the big Chinese one?

I forget.

There's always this weird random
statistic you see on searches.

But the, the point is like, for now,
these companies, OpenAI, Anthropic,

Google itself, you know, xAI, all, all
the various big, companies developing

large language models that virtually all
other AI, AI platforms use, they have to

live in the same world we live in, right?

They have to kind of figure out how
to optimize the AI to-- It's not just

about the world optimizing for AI.

They have to optimize the
AI for the world as well.

So by doing that, just by the sheer nature
of the internet at the moment, a lot of

this is gonna be very Google-focused.

there will be some adjustment
as that goes, and I think, it's

going to take more and more off of
Google being the dominant factor.

But they still have to use the roads
that were built by the Romans, right?

Like, they still have to use the--
they still have to use HTML and HTTPS.

And that's not to say that's gonna be it
forever, but for right now, and because

of the inertia that that provides,
structurally it'll look very similar,

even if five, 10, 15 years down the road
it is a new, you know, a new protocol, a

new, you know, markup language, whatever.

even if it is something brand new and not
Google's UCP or HTML or whatever, it'll

probably look relatively similar as far
as the nuts and bolts go, just 'cause

Brandon Giella: Yeah

Ian Andersen: that's where
you're starting from, you know?

And so the point with, with loyalty
program managers is, if these are the

generally accepted user protocols at
the moment, if we build to it, even

if it changes to something else,
you've done ninety, ninety-five

percent of the legwork, right?

So the switching costs will be much
cheaper than, than you might think.

It's not, oh, we're gonna invest
heavily in, in Betamax and then VHS

throws out all our effort, right?

It's, it's we're gonna invest
heavily in, you know…

I don't know.

I'm struggling to think of a good analogy.

But like option A and option B turns
out to be the winner, but they're

ninety-five percent the same, so it--
you don't lose any of that, that effort.

Brandon Giella: Yeah, I think what's
useful about a lot of these technologies

is they do build on each other.

You know,

Ian Andersen: Yes

Brandon Giella: MCPs are only useful
because APIs are useful, you know,

Ian Andersen: Exactly

Brandon Giella: you know,
that, that kind of idea.

It's

Ian Andersen: Yeah

Brandon Giella: uh, the economic
principle in, uh, of lock-in.

You're not locked in by

making this, this sunk cost, you know?

Ian Andersen: Exactly.

And that's, uh, I'm glad you brought up
APIs because l- that's another example

of, of there are other ways to transfer
information from one server to another,

from one website to another, whatever.

But over time, uh, the API framework
of, of, you know, standard curl requests

or, or, you know, the way that, that
you can do it in, in Python or, or,

Java have sort of organically centered
around, APIs as the primary method

for transferring, objects from, you
know, one software platform to another.

so if program managers can, can take
those basics and make it, API-ready, you

know, your, your balances, your, your
rewards balances, the, the earn rates,

redemption values, you know, tiers, y-
all of the stuff that, that is sort of

fundamental to loyalty programs across
the board, if you can structure that

data in a API-ready format, your, your
battle's nearly over at that point, right?

Then it doesn't really matter
which AI wins, which, you know,

commerce protocol wins, whatever.

You have your data ready to
plug into whatever system.

and so that's what, that's what I
mean by program managers having to

think through the, the, the paradigm
of how their platform is structured.

it can no longer be…

I mean, and it's been this way
for a while, even before the AI

conversation, but, but AI is, like,
really driving the lesson home.

It can no longer be totally native,
totally internal, totally proprietary in,

how things are structured just because
if your data isn't easily accessible

by somebody not typing in your website,
you know, and going directly to your

platform, then, you know, in a couple
years it might as well not be there.

And, and we have examples of this
from before AI in, airlines and hotels

when, when booking platforms such as
Switchfly, but, you know, Priceline,

Hopper, you know, whatever, all the
various booking platforms out there.

When the, when those started to,
take off and do, um, you know, and,

and consolidate, travel itineraries
for people, there were, there

were organizations that resisted
allowing Their data to be, be used

by, by those companies, right?

by, by booking platforms.

And whether it was airlines or hotels
or what, there were, companies that

did not allow their specific product
to be shown, and they took a hit

big time for not doing so, right?

Like it became readily apparent
that if you're not participating, if

you're not able to make it third-party
readable, you're gonna lose out.

And, and I believe AI is no different.

You know, if your data is not AI readable,
you know, back to the SEO conversation,

there's now something called AEO, right?

Answer engine optimization.

If you're not structuring your content
to be readable, easily readable

and understandable by AI, it's not
gonna bubble up to the surface.

So, it's no different for the
actual product you sell or the,

the rewards and loyalty, program.

If that information is not readable
by third parties, it's gonna go

by the wayside pretty quickly

Brandon Giella: Part of the point of
having a, an incredible travel podcast

is that you have content readable by AI

Ian Andersen: Yeah.

Brandon Giella: Anyway, uh,

Ian Andersen: Yeah, exactly

Brandon Giella: what, what you're
talking about, what comes to mind,

this might sound really esoteric.

Go with me for a second

Ian Andersen: I, I'm all about esoteric

Brandon Giella: What I know it, same.

Uh, w-what you're talking about, and
I can empathize with people, uh, not

wanting to get on the new paradigm.

and what I mean by that is back in the
day, this is before I was born, but

what I understood is you used to call
somebody to set up your travel itinerary.

They were called a travel
advisor an agent, and you would

talk to them, and they
would book things for you.

And really good ones had all this
incredible insider knowledge.

They had all this context about
places that they have themselves

been or booked trips through.

They knew these little
hole-in-the-wall restaurants or

hotels that you gotta go see.

If you go down this little path,
you get this incredible view, but

not many people know about it.

I'm gonna make sure that you have
time for that in your itinerary, etc.,

etc.

And then digital platforms came
online, '80s, '90s, 2000s, and those

travel agents then started using
digital platforms to do a lot of this.

But there was context that was lost
because you had to fit a protocol.

You had to get all of that knowledge
into a technical language that could

understand it and then present that
information and make decisions about it,

booking things, prices, dates, whatever.

And then, uh, then there are
self-serve digital booking platforms.

So now I, as an individual traveler,
can go to a Google, a Switchfly or

whatever and start booking things myself
with the options that are available.

But I have nowhere near the
context that a travel agent would.

And so there's

Ian Andersen: Sure.

Brandon Giella: loss as you're
increasingly more digital.

And what I'm hearing is same kind of
thing may be happening, where an AI

agent, in order for it to understand
the options or inventory available or

the payment methods available or the
scheduling or whatever, context loss that

a, a booking platform like Switchfly,
i-i-- that context gets lost, so it,

it, it has fewer options to book for
you because it has to be agent readable.

And so as you're going down the technology
chain, you're losing context, and I can

imagine would be people like a, like a
Socrates not wanting to write anything

down because the, the power of his word
or the depth of his word is, is lost

when you write it down, which is true.

Ian Andersen: Right

Brandon Giella: I, I don't wanna give
my data over in that way because I

want all of those options available.

It's very important that I ha--
retain all that context on my

platform because it's built a certain
way Is what I'm saying, is that, is that

tracking with how you're seeing things?

That because of this protocol, it's
limiting the-- there's context loss

from travel agent to digital platform
to agent to platform what's available

or what could be booked or, you
know, and there's some hesitation

maybe there with this new paradigm

Ian Andersen: Yeah.

I, I mean, you're not wrong in
that any time data is transport-

transferred between any medium, some
of the context is gonna be lost.

When our grandparents used to call
their local travel agent to, book, you

know, gr- they have a vacation they
wanna book or whatever, they have to

provide that context, whatever context
around the criteria of their, their

travel plans to that agent, right?

They have to say, "Hey, we have
this window available in the summer

that we'd like to go on a vacation.

we, you know, here's about
where our budget range is.

We're more beach people than
we are mountain people."

You know, they have to give that con-
that context to the agent, who then adds

their, their knowledge and expertise
around, you know, what you're saying, of

being able to then connect those dots.

If, you know, our grandparents
didn't tell the agent the timeframe,

they might come back with something
completely non-feasible, right?

Because that, that context
wasn't there at the beginning.

and I think it's the same at any point.

A, a good travel agent would go
back to pull the context from

the people go- calling them, the,
from our grandparents, right?

They would say, if they weren't
given it upfront, they would go

back to them and say, "Hey, when
are you thinking about traveling?

What are the kinds of
activities you like to do?"

The-- Right?

They would pull that, and that's where
that knowledge base that the travel agent

had is very important, 'cause not only do
they know where to go for the information,

but what information is pertinent, right?

And what could potentially
become stumbling blocks.

And I think we're at that point in, We
can plug in information to, uh, AI agent

to, to go book travel, and it still is
gonna require some of that context as far

as, you know, dates of travel, s- where
you're going, b- you know, whatever.

but it may or may not have that
built up knowledge to pool, what

else might be needed, right?

Brandon Giella: Hmm.

Ian Andersen: and I think where we
get around that or where that's going,

whether or not you want to get around
it or not, is as this stuff grows

and learns, it will be able-- it will
accumulate that knowledge base that, say,

the travel agent had of knowing where
to fill in those details of context.

for example, I think, right now, as of
right now, and I think everybody sort

of understands this intuitively but
may have not thought it through, if

you, tell an AI agent to go book travel
for this date range going to Tokyo,

you know, here's my budget, go, It's

gonna optimize for price, basically.

Uh, with, without any other external,
constraints around where and when

it should be looking and whatnot, it
will optimize for price, just because

price is the kind of fundamental
that we all have to deal with, right?

So, like, it's just a very good,
like, all else equal, we're

going for the cheapest option.

Brandon Giella: It's

quantitative, meaning a

Ian Andersen: Yes.

Brandon Giella: that out, whereas it

Ian Andersen: Exactly.

Brandon Giella: you know, for

Ian Andersen: E-exactly.

And the value judgment piece that's gonna
be brought in, that, say, individual

company might have, by the nature of their
business, that, that sort of knowledge

base that they've built up that, that
you're referring to, they might be

able to make certain value judgments
on behalf of the customer that an AI

wouldn't necessarily be able to right now.

So some of that's gonna be lost.

I mean, that is just the, the
nature of the world, right?

When, when you switch communication
mediums, some of that will be lost.

However

This is clearly the way the industry is
moving, and if you're not taking part

in it, it doesn't really matter how much
knowledge and context you have built up

if people aren't finding your product.

And this is the way people are gonna
be finding and booking products.

So that's the, the downside of it a
little bit is some of that will be lost.

It's just, it's gonna happen.

However, if you build your program
in such a manner to convey as much of

that as possible to an AI agent, there
will be possibility of that getting,

getting transferred through, right?

If, if, uh, you're structuring your
loyalty program correctly, if you're

structuring the linking posture of
your products, if you're, conveying

in a quantitative way the value of
your product, now the AI agent has

something else to, to take in and,
and will be able to better evaluate

considering those extra value judgments.

Um, the best example I can think of, and
I'm not sure if this is totally, a perfect

analogy, but like say you're a eco…

you run eco tours or whatever, that might
be slightly more expensive than a separate

tour that covers the same area but are not
like eco-focused kind of, kind of thing.

Well, your competitive edge, your, your
value prop that you've been selling on

for so many years is we are eco-focused.

You know, not only do we do tours
of the environment, but, you know,

our-- we have electronic vehicles
and, you know, and everything else

around it is we are eco-friendly.

if the difference on paper is only the
price, you're obviously losing out, right?

So how do you convey that message
to a, a computer that is built

to read quantitative, options?

It-- you have to find ways to, to
structure that in a quantitative way.

So it might be, and, and I'm
talking a little bit out of my depth

on the engineering side, right?

Like this is for very smart engineers
to figure out, not nerdy marketers.

But, if you can, if you can create
a, a quantitative value for what that

difference is, what that eco tour
difference is and, and offset it,

offset that from the price, right?

You now have kind of one-for-one value
that an AI agent can take into account.

is that, is that perfect?

No.

Uh, is it gonna be tough to do?

I mean, you said before our CEO all the
time brings up, you know, if, if everybody

could do it, everybody would do it, right?

This is complex stuff, that is gonna
take some real product leadership and

real engineering development, as well as,
you know, from the user side trying to

figure out the, the user side of this.

and the user side is really
interesting because you are now

optimizing for two users, right?

An AI agent and the end customer.

so it is, it is, it's not necessarily easy
stuff, but if the loyalty program is to

s- be successful, it, it needs to happen.

It needs to be done, right?

We are moving towards a more AI-involved
world in every aspect of our lives, and,

and booking travel is by no means exempt.

and even if, even if the, the agent
is not authorized to book to make

the final booking decision, the
options they present the user are

still gonna be relevant, right?

And if, if you're not even an option, then
that, that user can't even make the, the

decision of who to book through or with.

So,

Brandon Giella: I suppose the

Ian Andersen: it's tough

Brandon Giella: it is there's
the supply side, which is what

you've been talking about a lot.

You know, if you're

Ian Andersen: Yeah

Brandon Giella: a, a, you know,
a, a program that's supplying

experiences, hotels, and so on,
you've gotta supply that inventory

in a way that AI can read it.

And then there's the demand side
of a user saying, you know, "I,

I want to book these things."

And it's trying to make both of those
readable, but then also, factor in

judgment, whatever the context might be.

You know, it could be
dates, prices, values, etc.

But presumably, I g-- maybe to, maybe to,
to counterpoint what I was saying, but to

kind of take your, your thread one step
further, is if I'm talking to Claude for

two years, it's got tons of context on me.

Same thing with Siri AI.

So I just downloaded, uh, or
upgraded to the iOS twenty-seven

that came out this week, Siri AI
is supposed to be wonderful and can

do all kinds of things for you now.

Siri obviously has tons of
context on me, years of photos

and messages, and who knows?

And it's supposed to be
all private and so on.

But if I say, "Hey, Siri, book
me a hotel that does this…"

No, please don't.

Ah, I shouldn't have said that out loud.

Um, you know, "Book me a hotel, uh,
in for this, this, and this date," it

knows I have two kids and a wife, so I

Ian Andersen: Right

Brandon Giella: a large room that
can fit a crib, and, you know, it

knows that I need car seats and my
transportation and, and so on because

it knows all this context about me.

But then how does the supply side know
all of that context to surface in the

program that Siri can then read, and
then the booking agent in between those

two th-- You know, I-- It's a lot to

Ian Andersen: Yeah.

Brandon Giella: but,

Ian Andersen: It's a lot.

I

think there are buckets though
that you can do that in, right?

Of like family or not family,
you know, like higher end travel

versus eco-friendly travel.

You know, there are ways to
bucket that information that it

doesn't have to be precise to you,

Brandon Giella: Yeah

Ian Andersen: But it, it can
fit your broad demographic

Brandon Giella: Mm-hmm.

Mm-hmm.

Well, I feel like I've
learned something today.

Ian Andersen: Okay.

Brandon Giella: what, what do you, uh,
what's your, what's your closing takeaway?

What, what should people be
left with after this discussion?

'Cause I feel like there's so much there
that we could go on a million rabbit

holes, and I could talk to you for hours

Ian Andersen: Yeah.

Brandon Giella: and context and so on.

Ian Andersen: Yeah

Brandon Giella: what, uh, you
know, loyalty program leader

or product manager listening to
this, what do you say to them?

What encouragement or sympathy

Ian Andersen: I say that if, if
you're not thinking about this

now, you need to get on it, right?

Whatever,

Brandon Giella: Mm-hmm

Ian Andersen: steps you decide to go
down, however you plan to deal with,

AI commerce, you have to have a plan,

Brandon Giella: Yeah,

Ian Andersen: Even if your plan's to
ignore it, that is still a plan, right?

Brandon Giella: right

Ian Andersen: and, and if you're not
thinking how to structure, not just the

messaging, not just the offerings, the
rewards, you know, everything on that

side of things on the r- program, but
actually how the engineering nuts and

bolts of the loyalty platform is, is set
up, you need to start thinking about it

and, and making sure that assuming you
want to be AI readable, making sure,

uh, your, your platform is engineered
in such a way to be, to be AI readable.

So, that's the only takeaway
really is if you are not thinking

about this now, it's time to start

Brandon Giella: us, as the title
of this episode suggests, your

loyalty program is invisible to AI.

Ian Andersen: Yeah

Brandon Giella: you need to be visible.

Let's figure out a way to do

Ian Andersen: Absolutely

Brandon Giella: Well,
Ian, thank you so much.

Uh, uh,

Ian Andersen: Thank you

Brandon Giella: like I really
learned something today.

This is great.

Um, I

didn't even know that Google
could, uh, could book you

things, so you've enlightened me.

This is really great.

Ian Andersen: Yeah

Brandon Giella: I hope this has
been enlightening to our listeners,

and we will see you next time

Ian Andersen: Awesome.

Thanks, Brandon