Airbnb plans to use AI and VR to revolutionise travel once again
Disruption in the hotel market is set to continue as the industry upstart eyes future tech
The rise of Airbnb has been hard to miss. It has revolutionised travel by making it cheaper and easier to find somewhere to stay, or to find a unique place to spend the night, while letting people make money from their property when they don't need it.
These days, it's noteworthy when at least one keynote at a technology conference doesn't use the example of either Uber being the world's largest taxi company without owning any cars or Airbnb being the world's largest hotel firm without owning any rooms to illustrate the concept of disruption.
But as Uber's concessions to regulators and Facebook's very public issues demonstrate, these young upstarts are maturing they now yearn for acceptance within the very industries they once sought to disrupt.
Airbnb is the undisputed leader in peer-to-peer accommodation, but it needs to continue to evolve to chase new business and to appease competitors and regulators who are taking an increasing interest in its activities.
Earlier this year it detailed the biggest evolution of its platform in its short history in an effort to become more than just an accommodation platform, offering personalised recommendations for travel based on machine learning algorithms. The company is also planning 'tiered' levels of service, diversifying options for hosts and ensuring a guaranteed standard of comfort.
Tech is, therefore, key to creating new services and inspiring confidence in its platform.
Mike Curtis, head of engineering at Airbnb, joined from Facebook five years ago and leads a team of around 40 engineers and a 1,000-strong technology group that help deliver these changes.
"It's exciting because we see Airbnb as one of the only travel companies out there to develop tech to enable better travel," he tells IT Pro. "All we have to do was to get people to travel."
His team of engineers develop the code behind Airbnb's mobile apps and website, while a separate data science team help with the business decisions. Both are essential for Airbnb's personalisation plans.
"The intersection of those two [teams] is machine learning and using data to enable a better digital product experience," he adds.
"We see the problems for Airbnb as pretty unique. It's this huge community of travellers and massive number of hosts. Every traveller has different tastes and something might be completely different for someone else.""The challenge we face is how we deliver the right experience for the right traveller," says Curtis. "That's a significant [task for] the technology department."
Machine learning influences Airbnb's search results, which are built on its own stack. The ultimate aim is to deliver an end-to-end itinerary based on a user's personal history and people with similar tastes, recommending destinations and activities.
This is by no means a unique ambition, but it would transform Airbnb from an accommodation marketplace into a platform that can compete not just with the hotel industry, but more established travel sites like Expedia and Booking.com.
"We definitely want Airbnb to be a place to become inspired," says Curtis. "Our mobile apps show recommended destinations already and over time we think that once we begin to use the data we have, we will continue to get better at recommendation.
"We primarily use first-party data but we do have some social graph connections. The first-party data we have is great because it's a closed loop system - You search, book and then leave a review.
"The thing that's so fascinating about Machine Learning is that with a rules-based system you can only what's best for everyone, but with a trained system you can get something that is truly personalised."
Despite not owning any physical assets, Curtis' future projects will seek to help Airbnb expand from the digital world into the physical world, effectively allowing the platform to be a concierge service.
"Something I have a lot of energy for is being with our guests on every single step of the trip," he explains. "Our product is not the mobile apps, it's in the physical world. The concept of a concierge has been around there for a long time, but they've only been around for people who have the means. This is an area I think technology can help.
"Why should a seamless end-to-end travel experience not be available to everyone? We think that's something Airbnb should help enable."
Despite these efforts at machine learning, artificial intelligence (AI) is still in its infancy. Airbnb's other projects, including a trial of chatbots in customer care and for communications between host and guest, have yet to produce anything compelling.
"Chatbots just haven't crossed the threshold to being truly useful," says Curtis. "I think in the next few years these will become truly viable."
Virtual Reality (VR) and Augmented Reality (AR) technologies are also on the radar and Airbnb has held trials of 3D floor mapping. The idea is that this will boost customer satisfaction because they will get what they expect when they book.
Some traditional hotel and event venue businesses are doing this as well, but the sheer number of rooms and properties available on Airbnb means it's unlikely this will become widely adopted until the technology becomes easy and inexpensive enough for hosts to do it themselves.
"The concept is interesting," says Curtis, but admits there's nothing planned as part of the future roadmap just yet. "I think for us it [won't] be... until the quality is enough. At the moment you need high-end equipment."
Curtis is excited, however, by the idea of smart home applications. "Imagine checking into an Airbnb home and your smartphone can adjust all the smart home devices automatically," he says, highlighting yet another growth opportunity for the company over the coming years.
Of course, many people have never used Airbnb. Perhaps they're not ready to take a leap of faith, may prefer the familiarity of a hotel, or simply don't want to use someone else's home. However, some have simply been put off by horror stories about charges for cleaning or damage.
Curtis is more guarded about how tech could be used to resolve these sorts of disputes going forward but believes there's scope for innovation to make the hosting experience more efficient.
"We have an entire team building toolsets for our hosts to help them run a more successful hosting operation," he says. "We do have a resolution tool for small claims, price prediction. This is something we're always working on."
2022 State of the multi-cloud report
What are the biggest multi-cloud motivations for decision-makers, and what are the leading challengesFree Download
The Total Economic Impact™ of IBM robotic process automation
Cost savings and business benefits enabled by robotic process automationFree Download
Multi-cloud data integration for data leaders
A holistic data-fabric approach to multi-cloud integrationFree Download
MLOps and trustworthy AI for data leaders
A data fabric approach to MLOps and trustworthy AIFree Download