Eckoh's own TrainTracker™ service for National Rail Enquiries has been heralded as the UK's first and most ambitious automated speech service, giving all UK rail travellers immediate access to up-to-the minute train departure and arrival time information for direct train services between all 2,500 stations across the UK mainland.

A bit of history…
Speech technologies began to appear during the Cold War years, pioneered by technologists working for the American Department of defence, allegedly to support international phone tapping activities.  But the technologies weren't robust enough for commercial us; early applications proved disappointing and confidence in the technology flagged. 

The real breakthrough came in the early 1990's, when 'speaker independent' voice recognisers were developed, removing the need for any individual using the system to 'train' the recogniser to recognise their voice.   This, combined with the advent of cheap and plentiful computer processing power, meant that large vocabularies and the diversity of spoken language could now easily be catered for.

So, if the technology question was answered in the 90's, why hasn't speech been more readily adopted?  The answer is simple, and lies not in the technology, but in the way it's been used.   Poorly designed services that don't take into account the way people want to communicate and interact, have driven customers to abandon calls and bewail the lost option to talk to 'a real person'.

It's more than technology…
At Eckoh we overcome this by using conversation specialists, script writers and psychologists to develop services that complement rather than suppress natural human speech patterns and decision making behaviours.   They begin by assessing the process the company is trying to automate and look at how successfully it's working at present.  If the process is currently managed by a live-agent service, they'll listen to its calls and create a detailed map of the way callers behave; what questions they typically ask, how and in what order.  Even more importantly, they assess the motivation for the call - what does the caller really want to know and what's the difference between what they say and what they actually mean? 

Having developed working scenarios, our dialogue designers will test them in focus groups and small scale pilots before rolling them out in industrial strength deployments.  And, when monitoring the performance of the service in action, we still keep our focus on the customer rather than the technology.  After all, a word recognition success rate of 98% isn't worth much if 80% of callers are failing to complete their transactions. 
 
There's no doubt that the technology for large scale commercial deployments of speech is in place.  The big question for anyone thinking of adopting it is whether they have the specialist skills to design services that will delight customers rather than frustrate them.