top of page

Powerful (Prediction) Technology is Not Enough – Here’s Why

“It’s all about the experience”, a user

A truly intelligent typing experience, one in which your thoughts are seamlessly transformed into written text is part of a foreseeable future we are extremely passionate about and humbled to take part in creating. There are billions of man hours to be saved by introducing novel technology into the archaic typing mechanism that hasn't really changed since the first typewriter was invented 150 years ago.

We always believed that improving Lightkey’s prediction capabilities will increase the level of engagement by users and help them get the most out of it. As a result, with every new edition we kept pushing our own technological boundaries, and took pride in the ability to introduce better performance, higher precision and better keystroke prediction coverage.

But… as many users might tell you nowadays, developing powerful technology is not enough as it’s all about the ability to deliver superb (& personalized) user experience. One that the common user is likely to enjoy and get used to quickly.

But how do you do that?

It took us 11 editions (and many sleepless nights) to realize that having adaptive prediction capabilities is not enough… While using Lightkey sounds rather simple (hit the TAB for selection), getting used to using it and essentially change a deeply rooted routine may be slightly more complicated. The key insight was that beside the change in standard typing habits (frequent use of TAB inside words/sentences), we each have our own unique typing attributes such as typing speed and visual focus (screen/keyboard) that need to be taken into consideration.

In other words, if a user is focusing on the keyboard while they type it doesn’t really matter how accurate and time saving Lightkey’s predictions are – they would simply miss them. Similarly, if a user is typing very fast, offering predictions that only save very few keystrokes may be considered as less productive.

These insights inspired us to develop a novel “adaptive experience layer” designed to smooth the transition from standard-mechanical typing to predictive typing by combining both the visual and hearing senses. The top features now incorporated in Lightkey 12 include Dynamic Prediction Interactivity and Sound Assistance. We’ve also added an Interactive Personalization module to encourage users’ engagement.

#1 – Dynamic Prediction Interactivity

What is your typing speed? 20 words-per-minute, 40 WPM (the common average) or perhaps a lightning fast 100 WPM?

While some of us ‘fly’ on the keyboard, others take their time or simply don’t have the skill or the ability to go faster. Serving all with the same solution is not ideal and thus a relevant solution has to be dynamic.

Our latest edition is about making the prediction experience highly beneficial for all typing speeds. We’ve added an automated mechanism that detects your typing speed in real-time and adjusts Lightkey’s prediction interactivity accordingly – the faster you type, the more likely you will only be offered higher impact predictions.

In addition, there’s a new exciting opt-in feature which enables a unique prediction interactivity configuration specifically tailored to the needs of people with disabilities.

#2 – Sound Assistance

The physical separation between our PC keyboard and screen inevitably results in having users who are likely to miss Lightkey’s suggestions that show-up on their screens as they simply focus on their keyboard.

If we could only give them a gentle ‘tap on the shoulder’ to check their screens more often…

We decided to do this by playing a subtle sound notification that will signify the user when there’s a high-impact time & energy-saving prediction they might like to consider. As the user grows confidence in Lightkey, we anticipate that the sound cue will trigger a reflex-like TAB hit response, potentially even skipping the need for screen verification.

#3 – Personalization Module

The road toward a truly personalized experience includes learning curves, both on the user side and the technological side. While Lightkey utilizes learning capabilities to gradually gain confidence in predicting the user’s content, users gradually gain confidence in Lightkey's ability to assist them as the predictions become more accurate and relevant.

In most cases, users are likely to make good progress and adjust to the predictive experience of type & tab, each with their own pace. However, we wanted to facilitate a process that will assist users with a longer learning curve. Lightkey will identify these users and initiate a friendly personalization process encouraging the user to answer some questions, try certain features, practice more with Lightkey's tutorial and if necessary – submit a request for support.

The End Result – What to Expect?

The end result one should expect, as many of our users tell us after using Lightkey for a few weeks, is to simply forget that it’s there. As one user recently told us, it was only when he had to switch to an unsupported platform that he suddenly realized how much he missed it.

Nevertheless, trying to help new users visualize how their daily experience with Lightkey might feel, we usually describe a 2-step process: first, you gradually gain confidence in Lightkey’s ability to successfully predict you. Second, you find yourself successfully ‘predicting’ Lightkey’s predictions more and more often.

It shouldn't take long before you’ll find yourself typing the first 2-3 characters of the first word in a commonly used sentence and then just TAB all the way through the rest of it!

Next, you’ll know you have reached the highest level of this confidence-gain process when you start using the ‘wishful TAB’ feature more and more often (i.e. when you hit the TAB key before Lightkey’s suggestion pops up on the screen – essentially asking it to come forward).

A master skill saved for those who focus on the keyboard, is occasionally hitting the TAB key – with or without the sound assistance – knowing that Lightkey will just get it right, avoiding the screen verification… (I just did that with the word occasionally… and again just now 😊)



bottom of page