Generative artificial intelligence (gen AI)
is breathing new life into voice technology, prompting a frenetic exploration
of use cases and applications across the travel industry.
From voice assistants to a revamped Amazon
Alexa, they vary wildly, covering customer service and business operations
through to real-time translation and conversational booking.
Bret Taylor, chairman of the board for
OpenAI, the company behind ChatGPT, even believes voice could become more important than text chat in the
long run. Many travel technology companies seem to agree, with a
range of new tools now being developed—experimenting with features such as
emotional intelligence or cultural nuances.
But experts say their success will
ultimately be governed by customer demand. For example, Gen Z travelers dictate
many marketing trends we see today. With growing use of apps like WhatsApp, are
younger travelers likely to prefer talking over messaging?
Emotional
context
Most businesses, regardless of the industry
they reside in, look to use AI to become more efficient, as well as grow
revenue. Targeting the hospitality sector, last year German startup Onsai developed a voice solution for hotels,
promising them “round-the-clock service.”
The company says its voice feature natively
speaks 25 languages and can adapt to context and cultural nuances when
potential or existing customers call. It addresses one of hospitality’s biggest
challenges: global guest base meeting local staffing limitations. For example,
a Berlin hotel receiving calls in Spanish, Mandarin or Arabic can serve guests
in their preferred language at any time without multilingual staff across
various time zones.
Today, Onsai CCO Tobias Koehler is looking
at how systems will be able to detect emotional states, identifying frustrated
guests before they complain, recognizing VIP guests by voice and adapting tone
based on the conversation’s emotional context.
Subscribe to our newsletter below
“We already recognize if guests are hurried
or relaxed and adjust our responses accordingly,” he said.
Onsai’s data shows most phone reservations
made through AI-powered voice assistants are for the same day or within three
days.
“Guests want instant confirmation and
aren’t price sensitive. We’ve seen customers achieve a 30% or more higher
average daily rate (ADR) on voice channels compared to other booking channels,”
Koehler added.
Travel technology marketplace HBX Group,
meanwhile, recently launched an AI-powered voice channel for customer service in
collaboration with Ayesa.
RateGain,
which has just announced it is buying Sojern for $250 million, offers UNO
VIVA, which it claims is the first central reservation system-integrated AI
voice agent for hotels. The company says it solves a major pain point because
40% of voice reservations go unanswered. Additionally, it handles bookings,
modifications, cancellations and upsells across 30 languages.
Fritz Müller, head of Europe at RateGain,
told PhocusWire that voice AI has the potential to become the most
transformative interface in travel because it makes the experience
conversational and immediate.
“For hotels, that means a guest could call
at midnight in their own language and get accurate answers on room upgrades,
loyalty benefits or late checkouts without waiting on hold.”
In the corporate travel world, TravelPerk
proposes smoother hotel check-ins through “AI voice agent calls” that get in
touch with hotels 48 hours before arrival to confirm virtual credit card
payments, while Derbysoft’s new AI voice agents also call
ahead to confirm payment, booking details and request invoices after checkout.
The
next step
While voice paves the way for smoother
operations, online travel retailers are renewing efforts to leverage the
technology—technology that has significantly advanced since Skyscanner led the
charge a decade ago.
The metasearch site was one of the first
travel sector brands to introduce conversational search interfaces when it released a voice skill for Amazon’s Alexa in 2016.
Adoption then was slow because the market was not ready at the time, according
to Skyscanner chief product officer Piero
Sierra. “What’s different now is that with gen AI, voice can finally deliver on
its promise: real conversations with memory, context and multi-turn dialogue.
That is a game changer,” he said.
Skyscanner is now rolling out natural
language search in its app, which can be typed or spoken. “Voice is the next
logical step,” said Sierra.
Priceline also aims to make travel booking
as “simple as having a conversation” following a tie-up with OpenAI, and the
companies collaborated on updating the online travel agency’s gen AI chatbot, Penny.
Hotel booking platform HotelPlanner also
has a voice tool
that speaks 15 languages. The company said it used data from over 8 million
reservation calls to train the voice assistants.
Democratizer
of experiences
In the tours and activities sector, the
application of emerging voice technology is easier to picture.
“Real-time translation in live tours will
be a powerful lever of democratization,” said Arturo Moreno, chief data officer
at Civitatis,
an online platform for booking activities, day trips and guided tours. “Imagine
a Japanese traveler listening in Spanish to a local guide or vice versa. This
opens access to experiences that were once limited by language.”
Real-time translation in live tours will be a powerful lever of democratization.
Arturo Moreno, Civitatis
Further ahead, Civitatis is exploring
personalized real-time audio guides that adapt voice, tone and narrative style
to each traveler.
Platforms like Vox Solutions
are helping tour guides offer real-time translations as well.
The potential grows when interactive voice
experiences are combined with devices such as smart glasses, blending audio and
visual inputs.
“A traveler could walk through Barcelona,
ask about Gaudí’s architecture in their own language and have an AI guide not
only explain it but also book related tours on the spot,” said Brennen Bliss,
CEO and founder at travel and tourism marketing agency Propellic.
In-stay
potential
A common issue with emerging gen AI
platforms is that companies want to see the ROI. It’s no different with voice
assistants, and hotels, in particular, have long been trying to crack this nut,
probably ever since hotels began putting Alexa devices into their rooms.
Marriott was a launch customer for Alexa for Hospitality back
in 2018.
A lot has changed since then, according to
John Campbell, head of innovation and AI at digital agency Roast. He
argues that while text-to-speech and speech-to-text have been possible for
several years, it’s the fact that “conversations” are now possible,
as gen AI helps voice platforms answer back and have memory to recall past
conversations and preferences.
“In the past, conversational experiences
needed to be programmed one by one by developers, pre-thinking what the type of
conversations a user could have with a bot,” he said. “Now, LLM technology can
adapt to the guest as the conversation develops. Added to this, using RAG (retrieval-augmented
generation) allows the experience to be developed where a hotel’s website,
FAQs, menus and so on can all be ingested into the AI, and the AI finds the
response.”
In the U.K., Roast works with Warner
Hotels, where Campbell said there has been a lot of success with in-room Alexa
devices. “The number one thing that people want to know is how much they’ve
spent, and what’s their activity or itinerary for the day,” he said at the
recent Urban
Living Festival in London.
Warner Hotels tends to attract older guests
who are more used to using voice assistants like Alexa, Campbell said, having
skipped the messaging phase and gone straight into using voice as an input.
“You’ll get examples where you can see in the logs for a three-day period,
they’ll use it 50, 60 times a day,” he said.
In terms of proving ROI, Campbell said the
hotels can promote a last-minute spa offer on the device’s screen if there are
spaces available the property wants to fill.
“You can book a session at the spa, you can
book a boat trip down the Thames, you can book archery, shooting. What we’re
seeing is this additional revenue, sometimes it’s just about driving the
awareness of these additional services by seeing the screen and letting them
know about it,” he said.
“I like the idea of the AI being able to tap into
local guides to recommend things to do, according to the weather and time of
day. Being suggested to visit a botanical garden when it’s pouring down isn’t a
great suggestion.”
Meanwhile, he said Roast was in a pilot
with a well-known hotel brand in Italy to send out notifications to ask guests
if they were happy to skip cleaning, which would save the hotel €20 each time.
“ADR has a ceiling. You can only go up to a certain point, So I’m expecting
adoption to happen. It could happen quite fast,” he said.
However, for all this to work, he said
hotel systems must be sufficiently modernized for platform integrations,
including with services such as Alexa. Room request systems cannot rely on sticky
notes or disorganized WhatsApp groups.
“The challenge is that voice AI is only as
strong as the data it can access,” agrees RateGain’s Müller. “A voice agent linked only to a PMS can
confirm a reservation, but it cannot combine that with POS data for upsell
offers or CRM data for personalized loyalty perks.”
Texting
versus calling
For Michiel de Vor, co-founder and CEO of Runnr.ai, text
remains the key—for the time being.
“We follow the demand of the guest,” he
said. “And we see that reaching out to guests via WhatsApp is extremely
effective by looking at the reach and the engagement of the guests. Feedback
from hotels is that this actually reduces phone calls and emails.”
However, next year he said the
Netherlands-based company, which raised €1 million last year, wants to offer
alternatives, including voice.
We also have hotels giving us feedback that the quality of conversation from us is higher than from their employees—because we don’t make errors.
Michiel de Vor, Runnr.ai
“Our assumption is we will have all types
of inquiries when guests are calling. So, when they are calling, it’s quite
urgent. Otherwise, they’re not really going to call you.”
Making bookings via voice means the AI
assistant needs access to rate and inventory, which can be complex. As a
result, calls may be predominantly about changing or canceling a booking, or an
inquiry to add an ancillary or an extra bed.
De Vor says Runnr.ai’s technology works so
well enough that guests are often unaware it’s AI-powered. “Sometimes guests
ask: ‘How can you reply that fast?’ Or they’ll go to the front desk and say, ‘I
just spoke with somebody from your team.’ We also have hotels giving us
feedback that the quality of conversation from us is higher than from their
employees—because we don’t make errors.”
The renaissance of voice technology is
likely to grow in tandem with the rollout of gen AI, while agentic capabilities
look likely to be another theme underpinning growth.
Experts also point towards the rollout of
model context protocols as key, allowing AI agents to access necessary
information and execute tasks directly. In the past weeks there have been
updates from Sabre and Apaleo, which stand to further accelerate the
development of voice communication, as well as commerce.
“Voice creates that moment when guests
think: This hotel really knows me,” said Onsai’s Koehler.
