Natural Language – Timing Your Run to Perfection

The world’s most successful companies are not always the ones that truly pioneer. A clear example of this is Apple, universally renowned not for inventing, but perfecting. Apple has demonstrated a canny knack of identifying which technologies are truly ready for mass market adoption and packaging them to perfection.

Even so there have been issues; battery life has plagued the reputation of the iPhone 5 for instance, but still users persist with it. The Siri experience is very similar. The TV advert flagship of the Apple brand on phone, watch and tablet has certainly had its faults. In the beginning Siri was the subject of some humour as people found out silly things they could get it to say, but Apple once again timed their run to perfection. Even though not everything was perfect, consumers stuck with both Siri and the Apple brand, and this continued support – and ongoing use – actually improved Siri massively.

Indeed, Siri improved in two ways. Firstly, the truly huge number of users simply saying things into the microphone improved the accuracy of its speech recognition immensely. Known in the industry as ASR (Automated Speech Recognition), it converts the users’ utterance to a text string.

So, as more and more people spoke to Siri, it became better and better at getting the ASR conversion right and less people had to deal with the frustration of Siri not doing something because the text it generated was not what the user said. Apple customers stuck with the technology and were rewarded for their loyalty.

The second area in which Siri improved was capability, which encompasses three things; breadth of coverage, accuracy of task and intelligence.

Breadth of coverage is easy to understand. As more and more people spoke to Siri, Apple was able to learn more about what they wanted Siri to do and simply continued to add this functionality into the mix.

Accuracy of task, is slightly trickier because it is a subjective assessment of how well Siri can understand really complex instructions. Essentially, it boils down to how well Siri does at picking the right bits out and fitting them into the right ‘slots’ in a task. Take for instance arranging a meeting. Sure, you could start a linear process of “Siri – arrange a meeting for me”, “OK, when would you like your appointment for?” And so on. But much smarter to be able to say “Fix a meeting with Dave in San Jose on Tuesday afternoon at 3 at the Regus office to discuss Budgets”. This type of improvement came over time as the developers behind Siri improved what is called the NLU or Natural Language Understanding technology that sits under the hood.

Finally, intelligence, and here I mean artificial intelligence of course. This is one of the most complex ones to understand, but it’s important in terms of delivering what the user actually wants, their true intent. Here Apple started by understanding the “context” of the user. Context here has two meanings. Firstly and obviously, it refers to where the user was in the context of the dialogue with Siri (half way through booking that appointment for instance). Secondly it refers to the context of the device itself; where was it (GPS location), was the user” hands free” and driving, etc. For example, today a user can ask for a restaurant recommendation from Siri and it will only recommend ones which are on the route the current navigation is running if the user is running navigation, pretty intelligent.

The point is that when Siri was first launched, users had no knowledge of the functionality that was planned for Siri, but they were fault tolerant because the technology was good enough to get started with and they were rewarded richly for their loyalty.

Why is this important? Because today many enterprises are reluctant to extend beyond the four walls they are comfortable with to follow their customers onto the mobile platform and into the realm of voice. They cite the poor quality of ASR technology or the lack of intelligence of the solution as barriers. But in reality, these obstacles are self-made.

The smart enterprises of today are already working with mobile, speech and NLU technologies to time their run and gain the mindshare of the customer as the technology crosses the tipping-point and becomes a truly amazing customer experience.

Of course we can’t ignore the issue of cost either; Apple spent a fortune buying Siri and have continued to invest heavily in its ongoing development. But today it really is possible to generate Siri like experiences for enterprise customers in a realistic timeframe and with a realistic budget.

One company leading this charge is Artificial Solutions.

Artificial Solutions is helping large enterprises time their run into this new technological landscape and reap the rich rewards of higher Net Promoter Scores, improved customer retention and lower cost of customer acquisition.

Isn’t it time for you to start running, being left behind is never fun!

10 Things to Ask your Next Digital Employee White Paper

In this white paper we suggest some questions you need to ask when selecting your next digital employee.

View this White Paper

Andy, who lives with his family in the UK, is Chief Marketing & Strategy Officer at Artificial Solutions. A regular speaker at industry conferences and events, Andy delivers insight on the rise of AI, the challenges businesses face and the future of intelligent conversational applications.

Leave a Reply

Your email address will not be published. Required fields are marked *

[ Webinar]
[ Webinar]
[ Event ]
[ Event ]