By: Tim Noonan tim@timnoonan.com.au
This presentation will explore the applications of IVR technology for people with a variety of disabilities, as well as the older population.
Central to the presentation will be emphasis on using human factors research and principles to design intuitive and effective telephone-based services for all.
PhoneScript, IVR development technology developed in-house by the author for Royal Blind Society will be described and demonstrated, which has been optimised for the rapid prototyping and development of powerful telephone applications which are able to present a range of rich information sources to callers, through synthetic speech. Royal Blind Society is a not for profit service provider to people who are blind or vision impaired in New south Wales and the ACT.
Applications developed so far in the PhoneScript environment include:
JobPhone for presenting structured access to job vacancy advertisements from the mycareer.com.au website;
LibTel for browsing Royal Blind Society's braille and talking book catalogue and allowing online ordering; and
Today's News Now structured access to the full text of Fairfax newspapers.
Some of the unique features of the PhoneScript environment which were requested by users and service managers include:
A development environment optimised for text-to-speech IVR services (most existing platforms are recorded-message based);
Automatic processing of text through extensive PERL regular expressions, so as to dramatically improve pronunciations, and to convert written conventions into their spoken equivalents. Examples are Australian place names, names of politicians, British pronunciations, reading out complex currency values, intelligent reading of dates times and date ranges, identification and clear rendition of telephone numbers, identification of compound words and more;
Sophisticated acronym processing module which is able to identify (based on a large number of context rules) whether upper case words should be spoken or spelt out. Most speech synthesisers don't employ enough context information to do this job very well;
Full 'review' mode, allowing a caller to navigate a document by paragraph, sentence or word. Any word can be spelt out;
A set of menus for adjustment and personalisation of speech parameters including speed, volume, pitch and personality.
Three separate sets of voice parameters, one for menus/system messages, one for article/document reading and one for help messages. This provides increased navigation context and can increase comprehension of reading (listening);
Based around a very high-level scripting environment which hides the complexity of preparing text-to-speech buffers, telephony controls etc. This means that limited programming skills are required to tailor or fine-tune application user interface elements. Examples of some scripting commands are "hangup", "say", "spell" "saysubst" "title" "SayArticle" etc;
The scripting language facilitates automatic compliance with the Australian and New Zealand standard, with respect to standard key assignments and timeouts, but these can be easily overridden as required;
An intuitive 'talking keypad' approach to alphabetic entry, which complies to Appendix B of the standard;
Centred around a database-driven approach to data access, allowing a clean separation of back-end processing of source information, and front-end presentation of information to callers;
Currently this environment employs DTMF for user input, and the advantages/disadvantages of DTNMF vs. speech recognition for Royal Blind Society’s applications and customer-base will be explored in the talk. In particular, the challenge of successfully implementing speech recognition for older people will be discussed.
In the United States, accessibility legislation has pointed out some key issues which negatively impact accessibility of IVR services. Timeouts, the ability to connect with an operator at any time, and the ability to repeat information any number of times are some of the concerns raised. The final recommendations from the published rules for Public Law 508 will be explained and discussed. IVR Access issues for people who are deaf or hard-of-hearing will also be examined.
A recent initiative by the Australian Bankers Association (ABA) is the development of industry standards for four aspects of electronic banking – namely ATMs, EFTPOS, Internet Banking and Telephone Banking. The author is involved in these standards working parties and will briefly talk about progress on the telephone banking front.
The presentation will finish with brief coverage of the current Australian and New Zealand IVR standard (AS/NZS 4263) which presently doesn’t apply to speech recognition input. Is the time right for the Committee to start meeting to explore an update?
All Content Copyright (C) Copyright 2006 Tim Noonan
website by twpg