June 07, 2012

The personal assistant

Cvvzc8rmrjqnbv1pxakc


According to Norman Winarsky and Bill Mark, co-creators of Siri, the personal assistant is the next big thing in tech -- and they're right. As far back as 1996 Nicholas Negroponte wrote in his book "Being Digital" that the digital personal assistant would be coming. Curating and personalizing the web plus the ability to communicate with human language is the next technical challenge. The concept finally hit critical mass with the unveiling of Siri and people love the idea. Consumers may not know it yet, but the idea of a digital personal assistant is highly appealing.

How can I receive only the information I want to see? What can this product do for me? How can it improve upon what I have now?

The answer to all three is in personalization, a concept that has been around for some time now. Winarsky and Mark discuss Siri, the latest high-profile addition to the movement; however, there have been others.

Take Pandora radio, for example. Like a personal assistant, it asks you what you enjoy listening to and plays music similar to it. Should you dislike what it's currently playing, you can skip it and mark it as something you don't like. Gmail's labeling of emails as Important also falls under that. And if you look closely, you may find some form of a personal assistant built into a product you use every day.

The premise of a personal assistant is simple: it looks at incoming content, curates it to your explicit or implicit tastes it, knows what you want to see, and shows it to you. People are their own filters to the content they want and don't want to consume, but the personal assistant would be doing the brunt of the work for them.

Interestingly enough, Contactually is a product in that same wheelhouse. Our take on the personal assistant is narrowly focused on getting you connected with people you care about, answering the question, "who should I be emailing today?"

The perfect personal assistant is built on making suggestions and learning as it goes along.

  • Suggestions and learning - Steve Jobs once said "people don't know what they want until you show it to them." This couldn't be more true. The ideal digital personal assistant would not only show you the things that has already been established to be interesting to you, it would also be intelligent about picking things you may not know about yet. And of course, not every suggestion is expected to be spot on at first, but the point is to learn from user input/feedback and improve future suggestions.

  • Intelligent, context sensitive suggestions - Does it ask you follow up questions? Does it make suggestions based on location? Essentially, the personal assistant has to be able to zero in on results by using any kind of input, whether it be user, or location-based, or anything else.


Without the ability to learn and adapt suggestions to the user's changing taste, a personal assistant wouldn't be a personal assistant. But this is the hard part, and it wouldn't be a competitive market if it weren't.