In the last year, New York University Stern School of Business
dean of students Conor Grennan has become a leading consultant and voice on
social media regarding generative artificial intelligence.
Along with leading generative AI training for students and
faculty at NYU Stern, Grennan has worked with companies such as OpenAI, McKinsey, NASA, PwC and more, and he is co-host of the “AI Applied” podcast.
In a discussion with PhocusWire, Grennan shared some
thoughts on why travel is one of the most obvious use cases for generative AI
and how consumer facing travel brands should be capitalizing on this
opportunity. He also explained why using the technology internally – for
companies across any industry - is the easiest way to gain a strategic advantage.
The conversation has been edited for clarity and brevity.
Can you begin by sharing your thinking broadly on generative
AI?
It’s not a digital transformation. It’s a change management
thing.
There’s no learning curve to this. If you talk to it like a
human, you’re going to get phenomenal results. The problem is your brain
doesn’t allow you to talk to it like a human because your brain never talks to
a computer screen like a human.
So if I ask you, how should I be using ChatGPT, you would
say ...?
The problem with generative AI and the reason I teach
generative AI in the way I do is because teaching in terms of use cases doesn’t
really work.
Subscribe to our newsletter below
The brain really craves use cases because the brain thrives
on pattern, prediction, automation, things like that. And on the pattern front
it needs to know what it is replacing. When the brain sees ChatGPT it thinks, “I
know what this is, this is like Google.” The same way if you see red glowing
metal you know not to touch. You don’t have to think about it, your brain
automates that.
The analogy I use is that it’s more like electricity than a
use case. If you were to go back in time and introduce electricity, people
would say, how do you use it? And if I just try to give you use cases for
electricity, we’re never going to get anywhere. But if I introduce a light
bulb, people would know how to use it. So it’s sort of like electricity - generative
AI allows you to do what you are already doing but faster and better.
One year ago as it began to seem everyone was talking about generative
AI, there was so much speculation about how quickly we’d have implementation
and adoption. What’s your take on the speed of this?
AI in terms of people searching for things, it is going to
happen fairly quickly. One of the few use cases that people actually talk about with
something like ChatGPT are travel plans. … It’s one of the very few use cases I
see over and over and over again. So people lock onto it.
The thing that people go to right away is, hey I’m planning
a trip. So they go on ChatGPT and then “look how amazing this trip is.” If
that’s getting into sort of the zeitgeist, I think the adoption in travel is
going to be much quicker than the adoption around almost any other area of
productivity.
But generally I don’t think adoption of generative AI is
blindingly fast. To be honest … it’s pretty slow. Or if people say they are
using it - I don’t mean to be condescending - they are using it, but they are
just using it for small things like email. It’s 2% of what they could be using
it for.
How broadly do you believe generative AI is being used for
business purposes?
The statistics are overwhelming about how many people are
using generative AI – those statistics are wildly skewed. That is slightly
anecdotal, but you don’t have companies using generative AI in the way you
hear.
So when you hear, “At Fortune 500 companies 80% of people
are using generative AI,” it’s just skewed. What’s happening is that OpenAI can
track that here is someone from company XYZ, I guess that company is using it. But
those are individuals, those aren’t companies with a company-wide strategy.
And even those people using it – they are just scratching the
surface. Somebody is emailing and writing a few things, and people think this
is phenomenal.
Again, it’s not about use cases. It’s understanding what
people in your company are already good at doing and then amplifying and
augmenting that. Tell me what you do well. Tell me, marketing expert in
travel, what moves the needle? And then we can talk about how you are going to
augment that.
Speaking of marketing, one of the big questions
being asked in regard to travel is what happens to customer acquisition and marketing
strategy?
When we think about how AI grabs information, people will
still be going to brands like Expedia and Kayak and all these places to be
planning trips for sure.
It’s not about use cases. It’s understanding what people in your company are already good at doing and then amplifying and augmenting that.
Conor Grennan
So first of all, you have to have AI built into your
platform. If people go into, for example, Expedia and find a phenomenal AI
experience, there’s not a lot of reason then to go [to ChatGPT or another tool].
Remember plugins? Very early on [OpenAI co-founder and CEO]
Sam Altman said I’m not sure plugins are going to work … he said because I’m
not sure there’s a great product-market fit. He’s talking about his own product
here. The reason is I think people want GPT built into [the platforms they
already use]. People want to go to Kayak, for example, and have a GPT experience.
Rather than go to ChatGPT and request “now let’s use Kayak,” “now let’s use Turo.”
So I think number one is that you have to have a phenomenal
natural language experience in your platform. If companies can integrate it, that’s
going to be pretty powerful.
You already have the gold – the customer – just don’t lose
them.
And what about SEO in this new world of generative AI-powered,
fully query-driven search?
Nobody really knows what it’s going to look like in terms of
SEO. Anyone that tells you they know, they don’t know. We’re all
guessing.
But I would position it as we are the brand that gives you a
unique, customized, tailored-to-you experience. That would be the push from me.
On pure, how do we get people coming to us - keywords are
still going to be important.
Generative AI works a little like Google. Carvana
did this amazingly where they made a million different AI-generated, personalized,
animated videos for everyone that bought a car through them – it was
incredible. So I think personalization, landing pages, things like that that
answer a specific query. I know people that are creating literally 80,000 articles to
address any and every question they can answer – because you can do that now.
It’s not rocket since. But it becomes a numbers game. Can
you make bespoke content toward whatever people are asking for. That combined with once you get them to you, then make sure
it’s a great experience.
But don’t think of this as just a tool – because if you just
come up with a tool, it’s a commodity and everybody’s going to catch up.
So step one for a company is what?
This is my soapbox - train your people. You hired these
people because they are good, now imagine augmenting them by 47% above what they
do. That’s a strategic advantage. When you get everybody in your company
thinking about this, then it’s not just a tool, you are giving everyone in your
entire company an Ironman suit to go after these things.
Nobody really knows what it’s going to look like in terms of SEO. Anyone that tells you they know, they don’t know
Conor Grennan
That’s why I push training so hard. Once you get that and
get everybody really understanding what this looks like – that’s something that
strategically is a differentiator and other companies will have a very hard
time following.
If you can amp up your internal operations, that will show.
It’s like eating healthier and then on the exterior you just look healthier and
you look great in a bathing suit and all that kind of stuff. It’s what goes into your company.
If you want to really differentiate, you want to take the people
you hired because they are very good at what they do and imagine increasing how
well they do their job in terms of strategically, in terms of making decisions,
in terms of communicating, in terms of taking your non-value-add work down from
60% to 30% - which then leads to a lot more time for strategic planning.
This is phenomenal for strategic planning. The back burner
things. Let’s re-think this. You are now getting back 30% of everybody’s time
to focus on that, while everyone at other companies are fighting fires. Internal
is way more important than external. You have to do both, but when you are
talking about a long-term strategic advantage, that’s the differentiator.
Finally I’d like to get your take on AI agents.
I think they are coming very fast, but nobody knows what they
will look like.
There’s the device called Rabbit – nobody knows how it works.
I’m not sure of the efficacy of it, but how they say it works is you start
attaching your apps to this – then it is pulling behind the scenes. Forget about
SEO because your generative AI search goes direct to a site anyway. I don’t
know if it will go that way, but even the thought of it might mean leaning in
to this will be amazing. If agents will have their trusted sources, then that’s
what it will be.
I’m not sure it will be a separate device. That future is so
opaque, I don’t think anybody knows what a new device will look like. I really
don’t think Rabbit is going to be it. I mean, our phone does everything already
– why are we talking about something new?
For sure we start with the software and doing it through your
phone and then maybe if a device comes out that’s way easier to use, maybe.