Battle of digital assistants

Technology will soon make PAs redundant

Human resource managers use standardised tests to assess applicants for specific roles. For example, a personal assistant (PA) must be able to take dictation and respond to verbal orders. A PA also has to handle appointment scheduling and rescheduling, work flexible hours, learn about the boss’ quirks, and fade discreetly into the background as and when required. Now, a new crop of PAs check all these boxes and offer other competencies as well. What’s more, these PAs are omnipresent and they don’t need salaries, medical insurance or annual vacations. The voice-activated digital intelligent personal assistant (IPA) is now ubiquitous — every smartphone has one. Every digital major bundles its own IPA into its products. There are many independent IPAs as well.

These IPAs complete a variety tasks from checking the weather to tracking bills and from giving road directions to checking health parameters. Contrasting and comparing the features of popular IPAs such as Amazon’s Alexa, Apple’s Siri, Microsoft’s Cortana, and Google’s Assistant can lead to hot debates between users. The most popular open source IPA would be Mycroft. Facebook is also said to be testing its own IPA, which is yet to be launched, though Samsung announced earlier this week that it has delayed the launch of its Bixby voice assistant on its new Galaxy S8.

IPA features are a major selling point influencing choices of smartphones, laptop and car infotainment system. Many tasks performed by “software secretaries” seem trivial. But as with the old-style flesh and blood PA, delegating such tasks saves time and energy. Simply staying hands-free is also a great advantage in many circumstances. There have been huge improvements in voice processing and natural language understanding. These make it possible for an IPA to “learn” a specific user's accent and phraseology. In the old days, it was necessary to speak exact command phrases with a limited vocabulary. Nowadays, IPAs will decipher “Where can I get a drink?” as equivalent to “List nearest bars”.  IBM’s Watson (which is not an IPA) uses the same natural language processing techniques to discuss medical symptoms with doctors and to diagnose cancers. 

Another big factor driving IPA functionality has been generic advances in artificial intelligence (AI) programming and deep learning techniques. These help the IPA distinguish between the likely habits and preferences of users. IPAs leverage access to user information and location, and tap into online resources like weather or traffic conditions, radio/ video channels, news, stock prices, user schedules, retail prices, etc., to satisfy users.

Each well-known IPA has its own strengths. Google knows the user most intimately, which makes it easier for Google Assistant to make predictive guesses. Alexa is cloud-based and works through the Echo speaker to control home entertainment systems, drawing content from Amazon Prime. Siri mixes music requests, even processing exclusionary commands like, “Play any western classical music except Wagner”. Mycroft can manage home security. Most IPAs try to hook users into a proprietary ecosystem of compatible hardware. Siri for instance, works on iPhone, Mac, the AppleTV, the Apple Watch and even in a car, via CarPlay. Google works on Android phones and tablets, Chromecast, and in cars with Android Auto. Cortana works on smartphones (Android and Windows), tablets and laptops. All the IPAs are trying to integrate with third-party apps like Facebook Messenger to extend functionality; interoperability with other IPAs is also desirable. For example, Sundar Pichai’s Google Assistant may need to connect to Tim Cook’s Siri if the two CEOs are trying to coordinate schedules.

It's already apparent that a majority of mundane tasks can be performed by IPAs and that ongoing technological advances guarantee that functions, features and support for more languages will be added continuously. This has far-reaching implications: It will soon make an entire layer of low-level white-collar jobs redundant even as it improves productivity and efficiency. It will be interesting to watch the global economy adjust to these changes.