BTN's annual answer book for business travel managers everywhere. Added this year: travel risk management
Lola CEO Paul English talks:
After working on Lola for the consumer market, Lola founder
and CEO Paul English, who was founder and former chief technology officer of Kayak, pivoted this spring to a
focus on unmanaged and lightly managed small and midmarket travelers. The
hybrid artificial intelligence- and human-powered mobile travel assistant is
now tweaked for business travelers and debuts today. Lola is available directly
to travelers; those who download the app prior to 2018 will get VIP services
free of charge for one year. That includes hands-on human intervention when the
need arises. In a conversation with BTN editor-in-chief Elizabeth West, English
hinted that developments for midmarket managed travelers are likely around the
BTN: How did you
figure out that Lola needed to target the business travel market, rather than
the consumer market?
English: At Kayak,
the consumer travel site, about 25 percent of our travelers were road warriors,
the ones who take like 10 or 20 trips a year. We never built any features specifically
for them. At Lola, even with the first release, about 75 percent of our users
were business travelers. We thought we better get to know what they loved and
hated about travel and then build them a modern, sleek, fast, beautiful app
that would be like the best consumer app on their phone, but for business
BTN: So did you
get to know them? What did they say?
English: Oh yeah.
We had groups of road warriors show us what apps they use and how they travel.
It was particularly exciting for me to hear their frustrations. And the
20-something travelers were very different than the 40-something travelers. But
certain themes came through in what they disliked about travel and what they
valued about travel. First, they are very busy people and they want an app that
knows them. They are experts on their travel and they don't want an app
delivering choices and suggestions that they would never take or that their
company says is too expensive. They want their most relevant choices in a
channel they can trust. Also, supplier points programs were a priority for road
warriors. One of the reasons they put up with business travel is for the
points. So when they want to take their family on vacation, they can go for
free. If you book through Expedia or another travel site, you don't get your
points. We book direct through Sabre so they still get their points. Third,
there really is a rage issue around equipment delays, cancellations and other
travel disruptions. When these things happen, travelers want to talk to a
person who cares about their travel.
BTN: You are
serving the small and midsize market, and it sounds like you're going after
relatively unmanaged travelers right now. Tell me about that strategy.
English: Our VP
of product came from Facebook, where they use Concur. He showed me on one
screen on his phone [where the Concur booking tool] showed three violations for
one booking. Our goal is to give every hotel, every airline. We'll give you
what is in guidelines and what is out. Road warriors don't want a workflow
where their boss has to approve it. With an enterprise tool, you need more
guidelines. With small and midsize businesses you don't. You'll have lunch with
your boss the next day. You have to face them that you are booking outside the
system. That social pressure is a great thing in an SMB that you don't have in
the corporate where you have that central finance department. [The social
pressure] allows for more flexible product.
BTN: Is all of
Lola's content served by Sabre?
hotel, Sabre is primary because we want to book direct so the travelers can
earn their points. We also have longtail through Expedia. Hotels that don't
have Sabre—we can pick them up through one of the longtail providers. We will
add some others soon.
BTN: Tell me how
the AI is working behind the scenes.
English: Our code
generates an algorithm for every single traveler based on their click stream
and their purchase history. The algorithm learns what [the traveler] likes and
will surface those results to the user. We also do user clustering and
similarity models so that when a traveler doesn't have a history of going to,
say, Miami, we can still surface a hotel choice that is relevant to them. We
look at things like price history and sentiment data. We parse millions of
reviews [to determine] which words provide key information about the types of
hotels the traveler likes. For example, I like the Gansevoort hotel in New
York, so [Lola] would find hotels with similar key words and sentiment data on
a hotel in Miami.
BTN: But right
now, you are not actually having the user interface directly with the AI.
There's not a chatbot.
English: We are
using the chatbot internally. I'm not satisfied enough yet to inflict the
chatbot on road warriors, but that is coming. It's more important to nail the
technology first. Right now, the chatbot [pre-populates for the agent] what the
user says, and in some cases, the AI will say, "This is what the answer
should be," and the agent can just send it back over. But I'm waiting for
95 percent accuracy, and we are not there yet. No one is. If a chatbot
misunderstood me in the middle of a [travel disruption], I'm already angry at
the airline or the hotel. Lola doesn't want to make it worse. We are training
the bots, and the AI is parsing all the human input. To do it correctly, you
need a large corpus of text before the AI can accurately get responses.
BTN: What about
revenue. How is Lola making money?
revenue for us is going to be commissions from hotel, not so much from air. We
will introduce service fees for VIP after our first year. Our goal is to be
dramatically lower than anyone else. If you ever looked at Egencia's pricing
sheet, there are like 30 things they charge you for. We want an incredibly
simple pricing plan for service that we've not announced yet. Cheaper than
anyone else and simple. Most revenue will come from hotel commission. The
question is whether there will be a SaaS aspect or a per-user fee. We are not
announcing that yet.
BTN: Do you see
Lola ever powering other agencies?
English: If you
think of companies that have thousands of agents—we are in negotiations with at
least one of those companies right now. But our strategy is to scale up through
partners. Their agents will have our software, but it will always be the Lola
BTN: Will you
offer reporting or other policy overlays that will make Lola more amenable to
companies that want more controls or insights into traveler behavior?
English: Let's talk in December.
Correction: A previous version of this article misstated Paul English's role at Kayak as CEO. English's role at Kayak was chief technology officer.
Managed travel has been a laggard in the business intelligence space. Pi has had the advantage of...
In 2016, AirPlus lost 127,000 corporate cardholders and €12 million in profit, and it will register a...
It's been six months since Alaska Airlines merged with Virgin America, and "all the big milestones are...