Saturday, January 16, 2016

When Will Siri Make That Dinner Reservation for You?

romantic-dinner-for-twoEvery time I use Apple Siri or Microsoft Cortana, I’m frustrated by how little these services provide as the personal assistants they claim to be. Answering simple information questions is a start, but I expect a personal assistant to make an airline reservation, book a table at my favorite restaurant or even order a gift online that my wife would like. Why are these things so difficult?

This idea was first formalized about 15 years ago by Tim Berners-Lee, the man who (really) invented the World Wide Web. He called his dream the Semantic Web (or Web 3.0), meaning that it could learn and interpret the user context just as human personal assistants do. Natural language processing tools, such as Siri, need these capabilities to make them viable assistants.

Of course, you have to trust a personal assistant not to share user context and personal data with the wrong people. Yet I don’t see many users worrying about the wrong people accessing their smartphone personal text messages online or someone stealing their dinner reservation.

Certainly there is a risk, and everyone needs to make their own risk versus reward tradeoffs.In trying to understand why this huge entrepreneurial opportunity seems to be technically challenging or difficult to sell in the market, I have assembled the following list of five key requirements that are relevant to the realm of online virtual personal assistants:
  1. Easy and effective communication ability. In order to carry out requested tasks, a personal assistant has to first understand what is requested, both written and verbal, in the context and language of the requestor. According to a recent article, current tools still miss around 8 percent of spoken words and do very poorly on the context.

  2. Display an engaging and intelligent personality. All personal assistants realize that requestors have different moods and personalities, so they need to be respectful, courteous and sometimes assertive. I still find the current tools to be unintelligent, flat, boring and way behind the technology and marketing curve of what is possible today.

  3. Environment and context savvy. A good assistant constantly improves their knowledge of the world around the requestor, so they know quickly what is really being requested and what the acceptable outcomes might be in the requestor context. This is where all current tools fail -- by not honing a local database to learn from and frame each request.

  4. High value-added skills and productivity. Today’s personal assistants seem focused on expediting the Internet search process and simplifying text-intensive device control commands on your smartphone or tablet. The real value from personal assistants comes from being able to complete outside tasks like ordering products or making reservations.

  5. Provide convincing integrity, security and privacy. This seems to be a big hurdle in user acceptance, but I haven’t yet heard of Siri or Cortana sharing any personal search requests with outsiders. I believe the standards work now implemented for the exchange and protection of medical and personal health information should mitigate this concern.
According to Gartner, less than 40 percent of American consumers today use the free personal assistant services on their smartphones on a regular basis, and the best projections are that this could double by the end of 2016. To me, that suggests a major opportunity for additional products, free and fee, incorporating a new level of focus on the principles outlined above.

I do see a few startups edging into this space, including Assistant, HER and Alfred. To date, most of these are highly focused on one task, like trip planning or maintaining your calendar. Thus you would need dozens of these assistants, all with their own personality and limitations, to keep up with one human personal assistant. It’s no wonder that customers see no solution yet.

Hopefully you can now imagine all the fertile ground this opens for aspiring entrepreneurs. If you are looking for that magic million dollar idea, it may be time to build a plan around this one. But don’t wait too long, because the din for a virtual personal assistant on the Web and on your device is getting louder and louder. Catch the wave soon, or get drowned when it hits!

Marty Zwilling

*** First published on on 01/06/2016 ***


No comments:

Post a Comment