Browse Prior Art Database

Intelligent call routing and prioritisation using on-hold speech recognition Disclosure Number: IPCOM000124438D
Original Publication Date: 2005-Apr-20
Included in the Prior Art Database: 2005-Apr-20
Document File: 5 page(s) / 71K

Publishing Venue




This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 30% of the total text.

Page 1 of 5

Intelligent call routing and prioritisation using on -hold speech recognition

Field of the Disclosure

This disclosure relates to telephone systems, and particularly to telephone systems that include call holding and routing of held calls to agents.

Background of the Disclosure In the field of this disclosure it is common for telephone callers to be placed into a call holding queue within a telephony system before being transferred to an agent who can deal with their query. Often, businesses that employ this sort of telephone system have multiple groups of agents (otherwise known as skills), that can handle different types of queries. For example in a banking environment, callers that own a platinum card may be placed into a different call holding queue and then answered by agents that have special training in how to give the best customer service. Currently the call holding queue that a caller ends up in, and thus the skill that they end up speaking to, is typically based upon either the caller's calling line ID or on analysis on pieces of information input by the caller to a front end Interactive Voice Response (IVR) system.

Statement of Disclosure This disclosure introduces another metric that can be used to intelligently target callers into the most appropriate call holding queue. When callers are on hold, it is common for them to talk either to themselves or to colleagues and express their thoughts and feelings about being on hold. In known implementations, any useful data that could be inferred from the caller's speech is not captured. This disclosure proposes using a speech recognition engine to 'spot' words or phrases spoken by the caller and use the interpretation of these spoken words to re-queue or otherwise modify the caller's on-hold environment. In addition, the data collected from the caller could be attached to the call and then provided to the agent that answers the call.

Brief Description of the Drawing(s)

Two systems and methods for on-hold call handling incorporating this disclosure will now be described, by way of example only, with reference to the accompanying drawings FIG. 1 and FIG. 2.

Description of Preferred Embodiment(s)

A caller would call the contact centre as usual and after being initially queued would be placed on hold listening to holding music until a agent became available. In this embodiment a speaker independent speech recognition engine would be active whilst the caller remained on hold and would pick up any speech from the caller and interpret it. The interpretation would be via a speech recognition grammar (collection of expected words/phrases) targeted for the types of calls being received. Examples of grammar contents could include but are not limited to: comments about duration of hold - "I've been holding for half an hour", queue direction - "I hope this is the right number to renew my policy", expressions of dissatisfaction - "This letter I've received is disgusting". Matches within the...