Finding a unique voice in travel technology

The increasing and widespread adoption of voice technology is already beginning to establish its place as a major factor in travel.

A truly intelligent voice technology platform would enable fluid, human-like conversations. Artificial intelligence (A.I.) technology would support complex exchanges, detect nuances in speech patterns and tone and remember details from earlier conversations.

As these become more widely used, travelers will expect to interact with voice interfaces to automate aspects of travel logistics.

Imagine asking Alexa to look up flight information and book airfare and hotels, or even calling a customer service desk and speaking to an A.I.-powered virtual agent.

Many of these scenarios are being implemented in the travel industry, but there’s still a significant amount of room for improvement.

Where are we now?

Voice has already changed the way we think about travel. Amazon currently lists over 830 "travel and transportation" skills available on Alexa, including major services like Uber, Lyft, Kayak and Expedia. But these are far from perfect.

Voice interface tools are still limited by their ability to accurately understand queries and by their limitations in retaining context.

Despite claims of voice recognition having equalled or surpassed human listening accuracy, real world experience is often very different. Things like background noises, accents and a host of other challenges still limit actual listening accuracy.

While voice technology still needs work to become nearly indistinguishable from a person, there are other inherent issues with shopping for travel via voice-enabled platforms.

Also, speaking is sometimes slower than typing, and listening is always slower than reading.

Combine that with the lack of a graphic interface and it makes things like shopping very difficult, especially in travel where you need a lot of information at the same time.

But what if travelers could view all the information while still controlling the shopping process with their voice?

Coupling voice with a graphic interface instantly makes the experience better. For example, we did a prototype experimenting with voice filtering combined with a screen on the [Amazon] Echo Show, and this method really shows a lot of promise.

The importance of security

Security is an important consideration for digital technologies, which includes voice interfaces.

Using voice tech in public places could be problematic, especially when dealing with sensitive information in shopping for travel.

Usually these conversations are more sensitive on one end. Changing flight plans on a headset while walking through an airport could be benign for the person requesting, but the content in the response might be private.

Earbuds might address privacy issues with these conversations, but don’t necessarily ensure security.

Another concern is that it isn’t really practical to password protect voice interfaces. This would require privacy, which isn’t always available with voice.

Voice works best when it’s available anywhere and retains context of who is talking and where they are at the moment.

So, there’s a huge privacy tradeoff for using voice, which may be a concern especially in some regions. This is something we have to think about as we develop this technology.

Where we’re going

While most people think about voice-based conversational interfaces will impact consumer experiences, another important factor is how these tools can improve business operations.

We think voice will be a bigger deal for operations than it will be as a consumer-facing sales tool, especially for maintenance crews, cleaning crews and similar operations.

Wearing an ear bud that allows you to receive important information or instructions while also allowing you to dictate notes and update maintenance records, that kind of hands-free interface can be extremely beneficial over a traditional screen.

Another operational area where voice technology will be prominent is in customer service and call centers.

The ability to log and analyze mood, tone, pace and other attributes of the customer interaction can significantly improve customer service and provide feedback to employees.

Additionally, conversational interfaces can automate routine requests, thereby freeing live agents to focus on more complex interactions and revenue-generating activities.

This might be a reach, but we would love to see "connectedness". In the future could there be a connected user profile or some sort of communication between third-party apps to improve user experience across the board?

At the rate we’re going, the answer may be coming sooner than we think.