Why are virtual assistant apps so shy?

Why are virtual assistant apps so shy?

The industry has been promising apps that interrupt us with important contextual information. So where are they?

We've been on the brink of a new relationship with our gadgets for years.

"Real soon now," we're told, mobile apps will interrupt us with personalized information we never asked for. These interruptions would show us opportunities ("Those shirts you like are half-off at the store around the corner"), facilitate our social lives ("Your friend Janet is also visiting New York and is free for lunch -- would you like to invite her?") and save our bacon ("You haven't ordered a flower delivery for your anniversary -- would you like me to take care of it?").

What are they waiting for?

Apple's Siri responds impressively but doesn't take the initiative to interrupt much.

MindMeld is supposed to ship to general availability an app that churns out information related to your video conversations. The technology and platform look impressive, but the purpose of MindMeld's proof-of-concept app will be context for conversations, not mobile personal assistance.

EasilyDo is supposed to give you predictive results, but like most in this category, you have to work for it. The results are a fraction of what you want, actually harvesting data from social networks, calendars and location-oriented databases and placing them in a stream.

EasilyDo performs a wide variety of useful tasks like contacting people for you, telling you when to leave for your meeting, getting directions, automatically creating new contacts and many other things. EasilyDo automates tasks to simulate a personal assistant, but it doesn't do anything you haven't specifically commanded it to do. It doesn't interrupt you proactively with contextual information beyond things like telling you to leave for your next meeting.

Osito is another cool app that suffers from the same lack of creative interruption as EasilyDo. It's more of a rollup of the kinds of information you can get elsewhere, but presented in a stream with some pop-ups. It's more about automation than contextual assistance.

Foursquare just got interesting. The company rolled out a test version of Foursquare that throws advice and context at you when you walk into a restaurant or around a neighborhood. ("Try the cheese fries in this place -- they're incredible.")

Checking in isn't required. In fact, the information will pop up even if the Foursquare app isn't running. Now that's what I'm talking about.

Sadly, the test is for 2,000 lucky Android users only, with a wider rollout promised for an unspecified future. Foursquare tested a comparable feature called "Radar" two years ago, but killed it after "Radar" killed battery life.

Google Now has by far the best proactive interruption in the business, if you have an Android device.

Google Now grabs information about you from Gmail, Google Search and elsewhere, and uses that data to improve results.

Lately, Google has folded in some amazing capabilities. For example, it preemptively feeds you information about your car rentals, public transportation information based on guesses about where you might want to go, movie tickets and sports scores.

Google Now tells you more information about whatever's on screen. It knows what you're watching because, with your permission, it listens to the sound of the show to figure out what you're watching.

Google Now is great in every way except one -- it doesn't give you enough information. Leaving Google Now running gives you a mostly static view, with "cards" coming in very infrequently.

Field Trip, another Google product, is wonderful only because of what it promises, not what it delivers.

We learned recently that Field Trip was originally designed for Google Glass but shipped on the iPhone while Glass was still in early development. It's now available on Glass.

Field Trip has the right idea, popping up contextual data. However, these are based on a list of arbitrary database-oriented websites, such as "Historic Detroit," "Public Art Archive" and " San Francisco Architectural Heritage." Unless you're in a major city, Field Trip contextual information is slim pickins.

There are other services that claim to interrupt with contextual information. But in my experience these require launching the app and refreshing it -- a conscious choice followed by deliberate action, which is the opposite of what's promised in this category -- interruption.

Looking at these few examples, the industry overall seems hesitant, unwilling or (most likely) unable to meet the promise of interruption for contextual information.

It's true that this category is fraught with hazard. People feel their privacy is being violated when a gadget demonstrates what it knows about them. It's an irrational concern because not letting the user take full advantage of harvested data (as is the case today for most users) doesn't equal privacy, just ignorance.

Users might get overwhelmed by too much interruption data. But there are better ways to stop overload than simply withholding messages. Let it be turned up gradually. Offer a snooze feature. Set times of day. Notify only when moving.

The ideal service would combine the benefits of Foursquare, Google Now and Field Trip. It would give you useful and interesting data not just for general areas or historic landmarks, but down to the specific address level.

It would be like the test version of Foursquare, interrupting you whether the app is running or not.

It would be like Google Now with its uncanny ability to know all about you and learn from your actions.

It would be like Field Trip in bringing in disparate human-curated databases to augment your reality.

Most of all, I would like these services to turn up the fire hose. Limiting access to a handful of users like Foursquare, dribbling out frustrating little bits of data like Google Now, and limiting information sources to a handful of content sources like Field Trip isn't going to cut it.

This category has been full of promise and potential for two years now. The wearable computing category is exploding with activity. It's time for the industry to get over its shyness and start interrupting us with more and better personalized contextual information.

This article, Why are virtual assistant apps so shy? , was originally published at

Mike Elgan writes about technology and tech culture. Contact and learn more about Mike at You can also see more articles by Mike Elgan on

Read more about emerging technologies in Computerworld's Emerging Technologies Topic Center.

Follow Us

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

Tags Applesoftwareapplicationshardware systemsmobile appsEmerging Technologies

Show Comments