On Talk Shows and Talking to Your Smartphone: Why Voice Will be the New Text

Siri AI Agent Last week I was asked to be a guest on Monday’s edition of AirTalk on KPCC – an NPR affiliate radio show based out of L.A. – to talk Artificial Intelligence (AI) and voice technology as the next consumer frontier. I wrote a piece with a similar theme a couple months back on virtual assistants and the “AI Arms Race.” Due to a scheduling conflict, I ultimately didn’t end up being on the show, but the topic is fresh in my mind and one I have a lot to say about.

I took a listen to the show on Monday and I enjoyed it. But there were a couple of points I disagreed on, and a few areas where I might have elaborated further. The current market of Artificial Intelligence agents is crowded, fragmented, and not particularly differentiated. It is nonetheless a huge market, representing roughly 70% of U.S. adults, owning over 200 million smartphones. The key players in this space (Apple, Google, etc.) have market caps of several hundreds of billions of dollars each. That’s not to say that there’s much money to be made in virtual assistants like Siri at their current stage. At the moment, these virtual assistants are not typically the deciding factor in whether consumers purchase one smartphone or another, and they don’t cost the consumer to use beyond the price of the phone.

But voice technology on smartphones is only the first step; there’s a broader consumer ecosystem at play. These virtual assistants want to be in your car, in your home, and even on your body. If Apple, for example, can get people comfortable and familiar with using Siri on their phones, they will be more likely to use Siri in their homes and in their cars as well – environments that require a slightly more deliberate behavior change, and potentially more trust in the reliability and security of the solution.

That’s where the real opportunity lies, and is why companies like Google, who in October just released its flagship Pixel smartphone with Google Assistant built in, are continuing to invest in this market despite Siri being available on the iPhone since 2011. In the 5-6 years that Siri has existed, it has not become the dominant way in which people use their smartphones. But it has huge potential. Studies show people check their phones 50-100 times per day (probably more for us Millennials). If Artificial Intelligence agents can replace the majority, or even some, of those touchpoints with contextualized requests, this will create an entirely new way of interacting with technology.

Guest, David Pierce, Senior writer at WIRED, said AI won’t replace your smartphone for certain things – such as when you need to read something or watch a video, or even book a flight. Now, no disrespect to Mr. Pierce, but that feels shortsighted to me. Can we not imagine a more immersive medium than text or video that may soon be the dominant standard, and that may include, in some form, AI technology? I can. Tech-fluent users will always find the easiest and most effective way to use technology, and as we’ve seen with video beginning to replace text, and online almost completely rendering print a thing of the past, the dominant standard for consuming media is in no way set in stone.

At its current iteration, sure, AI technology is not well suited to performing a general range of tasks. It is optimized to perform a highly specialized and narrow task, such as setting a timer or playing a song. But Pierce’s example of booking a flight is actually not far off. This kind of personal assistant task, traditionally performed by someone like a travel agent, and now done almost exclusively online, is exactly the kind of duty a virtual assistant will quickly become optimized to perform. Booking a flight involves a logical and routine set of responsibilities. Where to? What timeframe – AM or PM? Do you prioritize timing or cost? Pierce describes an AI interaction in which the virtual assistant simply lists out every available flight in some kind of droning monotone. But with a few programming tweaks, I could imagine the AI assistant asking questions and finding you the flight you are looking for in quite a pleasant exchange. I, for one, will not particularly miss the endless scrolling and filtering of traditional online flight searching.

John McCarthy, a computer and cognitive scientist who was one of the founders of Artificial Intelligence as a discipline, and who coined the term Artificial Intelligence, had a great quote: “As soon as it works, no one calls it AI anymore.” There’s a lot of progress to be made with AI, especially in its ability to understand colloquial language and slang, and expansion beyond U.S. English as the default language. Much of this requires software improvements. But hardware also needs to be updated, especially microphone technology that has largely remained unchanged for the past 5 years or so. Unsurprisingly, if the virtual assistant can’t hear you, it can’t understand you either.

Context and technology need to come together in order for the value of AI agents to exceed the roughly 30 second time savings produced by using voice to set a timer, and tasks of that nature. This will take at least 3-5 years – in replacing the current generation of hardware, and enhancing the software to understand, not just the literal definition of words and phrases, but also the subtler language nuances behind them. Additionally, it may take a few years for consumers to overcome their initial disappointment with prevailing technologies like Siri and Amazon’s Echo, that haven’t yet elevated voice to the same level of usability as touch interaction. Consumers will need to begin to trust natural language interaction as an efficient means of performing routine tasks they have already become proficient in using their smartphones for.

Next Steps for AI As for what AI will look like in the future, there are endless opportunities. Most seem unbelievable: some unbelievably good, others actually quite frightening. Though it doesn’t appear this way from the current iteration of narrow-application virtual assistants on smartphones, AI, once it reaches general-, and (gulp) if it gets to super-intelligence, will fundamentally change every facet of our current lives. David Pierce jokes it will be like the TV show, Westworld. I always imagined it being a little less Western.

About Charlotte O'Donnelly

Charlotte O'Donnelly is a Research Associate at Blue Hill Research supporting written and research topics in mobility, IoT, and technology expense management. She is primarily responsible for surveying the market and reporting on significant trends and developments from market leaders in this space. Charlotte also supports the analysis, writing, and creation of client deliverables, multimedia assets, and internal initiatives. Prior to Blue Hill Research, Charlotte worked in mobile technology and financial services consulting. Charlotte has a background in business, technology, and law, and is passionate about the intersection of these subject areas.
Posted on December 7, 2016 by Charlotte O'Donnelly

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Latest Blog

Blue Hill Research Communications Lifecycle Management Highlights: April 2017 This Week in DataOps: Rain, the Real World, and Another Manifesto (the Good Kind) Network and Device Security in the Age of IoT

Topics of Interest

Advanced Analytics

AI

Analytics

Anodot

Attunity

authentication

BI

Big Data

Blog

Business Intelligence

Cloud

Cognitive Computing

Corporate Payments

Data Management

Data Preparation

Data Wrangling

DataKitchen

DataOps

DataRobot

design

design thinking

Domo

Emerging Tech

enterprise applications

Enterprise Performance Management

enterprise video

fog computing

General Industry

GoodData

GRC

Hadoop World

Human Resources

IBM

IBM Interconnect

Iguazio

ILTACON

Informatica

Information Builders

innovation

Internet of Things

IoT

knowledge

legacy IT

Legal

Legal Tech

Log Data

Machine Learning

Managed Mobility Services

Microsoft

Mobility

Nexla

Order-to-Cash

passwords

Pentaho

Podcast

Predictive Analytics

Private Equity

Procure-to-Pay

Qubole

Questioning Authority

Recurring Revenue

Risk Management

ROI

Sales Enablement

Salesforce

Security

service desk

Social Media

Strata

Striim

Supply Chain Finance

Switchboard Software

Tableau

Talend

Tangoe

Telecom Expense Management

Time-to-Value

Trifacta

TWIDO

Unified Communications

usability

USER Applications

User Experience

User Interface

video platform

Virtualization

Visualization

Wearable Tech

Yellowfin