Home » Services » Examples of adjustments and discrimination: service providers » Voice recognition telephone systems » Voice recognition telephone systems: Archive September 2012

Voice recognition telephone systems: Archive September 2012

Disclaimer – please read
This page does not apply outside Great Britain.
This is an archive of the page as at 25th September 2012, see explanation below.

This is an old version of the page, for the record since it was current at the time a parliamentary written question was asked. The present version of this page is here.

Voice recognition systems are one of the biggest barriers to phone communication for people who stammer. They may involve a breach of the duty to make reasonable adjustments. There should be an option for prompt transfer to a real person. For a summary on accessibility for customers who stammer, see Making services accessible.

Telephone systems which understand speech are becoming increasingly common. However, voice recognition software is unlikely to understand a person with a significant stammer or other speech difficulty. Accordingly there needs to be an option to get through promptly to a real person if the computerised system cannot cope.

A reasonable adjustment might be – I suggest – an option to speak to a real person given either in the initial menu, or the first time the computer fails to hear or understand the caller.

The option to speak to a person should of course be available through pressing a key, rather than being invited to say e.g. “operator” which the person who stammers may be unable to do.

One of the ‘top tips’ in an Ofcom document on call centres is that an option of speaking to an operator should be included in the initial menu. See Disabled customers and call centres (link to archive of ofcom.org.uk).

Possible alternative adjustments:

  • saying the caller can simply hold to be transferred (promptly) to an operator; or
  • allowing the caller to select an option by pressing a particular key as an alternative to saying something.

Case studies

Case study: A woman who stammers phoned her local hospital’s switchboard number, to speak to a friend who was a patient. The hospital had converted its system to voice recognition. She was asked to say the name of the person or ward she wanted to contact, or say “operator” to speak to an operator. She was unable to say any of this and waited till the computer spoke again to see if there would be a push-button option, but there was not. She put the phone down, unable to get through.

Case study: Someone who stammers tried to book a train ticket by phone. The system used voice recognition. He had trouble with b’s and could not say ‘book’. He was switched to a real person only after failing about a dozen times to be understood by the computer.
(Clearly this is too long: it should be possible to get through to a real person much more quickly.)

Case study: A person phoned a call centre to check that a letter sent by recorded delivery had arrived okay. He was required to read out a series of letters and numbers from his receipt. It took several tries before he managed to get the computer to recognise what he said. (Some who stammer would not succeed at all.)
See Say it again, says telephone Sam (archive of stammering.org), 2006.

Normally, the solution would be to build in the choice to speak to a real person. However, exceptionally it may be possible to make special arrangements for an individual who stammers:

Case study: A school installed a new system for recording absence. This asked for the parent’s name, child’s name, class and reason. A parent with a stammer emailed the school and explained the difficulties she experienced. The school agreed to let her email the absence officer direct.

20th anniversary of stammeringlaw, 1999-2019