Building apps for the $30 smartwatch
Gartner predicts that the smartwatch market will take off in 2015, driven partly by Chinese Android imports that will bring the price down to just $30. Whilst there will still be manufacturers able to charge more by offering additional value to the product, at that kind of price point the device is just a conduit, the money to be made by other companies will be through apps and services, but do they have the skills to do it?
Obviously voice and intelligent understanding will be key to interacting with a smartwatch, nobody’s fingers are small enough to start scrolling through menus. However, it is this aspect that will prove the most difficult to achieve as users start to expect more functionality from their wearable devices. Speaking to an app, that’s easy, but as anyone that has already tried to develop complex applications that use natural language beyond simple commands will know, not all natural language platforms are equal.
There is a reason it has taken so long for natural language to come to the fore and surprisingly it wasn’t because we were just waiting for Apple to release Siri. Computational linguistics, which is at the heart of natural language, is complicated enough to make a rocket scientist go cross eyed.
Most natural language platforms still rely heavily on computational linguistics in its rawest form to develop and maintain applications. For organisations this can create several problems and not just the expense of maintaining a team of experts skilled in what some people describe as a black art.
For UX designers it can mean that they need to be able to physically draw up all the variables of how a question might be phrased in order for them to be programmed. Business process designers don’t have visibility whilst the app is being built and so can’t point out any misunderstandings until it’s finished. And that great idea for a new feature that the sales director had once he saw the app? – well forget about it until the next release – in six months time (hey, we all know it will nearer eighteen, after all this is software).
Sounds painful doesn’t it?
It’s one of the reasons that five years ago Artificial Solutions decided to turn the world on its head and put the business user first in the development of natural language applications with the release of Teneo. We’ve come a long way since then. Such as allowing users to add an answer to a dialogue by just give two examples as to how the question might be phrased to give it some context and Teneo works the rest out for itself.
We’ve packaged the complex lines of coding into easy to visualize flows, so the person that deals with that part of the business everyday can see instantly if there are any errors in how a process is handled. What’s more adding features, languages, etc doesn’t mean going back to beginning and reprogramming the entire app. Teneo has a modular structure.
So does that mean computational linguistic engineers are on their way out? No way. At Artificial Solutions we have some of the greatest minds in computational linguistics working for us. But that’s just the point, it’s a resources that most app developers can’t afford.
Natural language is all about intelligent interaction with machines; shouldn’t it apply to the way you build the app too?