System and method for providing an automated call center inline architecture

1. A system for providing an automated call center inline architecture, comprising:

a script engine to maintain prompts and a plurality of grammar references;

a telephony interface to receive a call and to collect audio data using the prompts from the script engine, which are transmitted to the telephony interface via a message server;

a speech server to perform distributed speech recognition, comprising:

one or more secondary recognizers to receive the grammar references from the script engine via the message server and to determine speech results by applying the grammar references to the audio data; and

a main recognizer to form a new grammar from the speech results and to identify speech recognition results by applying the new grammar to the audio data;

a confidence level module to assign confidence levels to each of the speech recognition results as an indication of accuracy of the distributed speech recognition;

a placeholder module to assign a placeholder to at least a portion of the speech recognition results when the assigned confidence level for those results do not satisfy a predetermined threshold;

an action module to perform an action based on a presence of the placeholder, wherein the action comprises querying the message server for additional speech recognition results; and

an agent console to receive the speech recognition results and the placeholder as a display.

2. A system according to claim 1 , wherein the action further comprises modifying the speech recognition results based on instructions from a user.

3. A system according to claim 1 , wherein the secondary recognizers simultaneously process and determine the speech results by applying the grammar references to the audio data.

4. A system according to claim 1 , further comprising:

a notification module to transmit a text message to the message server upon the receipt of the call through the telephony interface.

5. A system according to claim 1 , wherein the grammar references comprise at least one of independent grammars and non-overlapping sections of a main grammar.

6. A system according to claim 1 , wherein the secondary recognizers apply an n-best algorithm to the grammar references and the audio data to determine the parts of speech most likely to occur in the audio data.

7. A system according to claim 1 , wherein the script engine maintains a main grammar and the main recognizer reforms the main grammar by combining the new grammar with the main grammar.

8. A system according to claim 1 , further comprising:

a response module to generate a response to the caller based on the speech recognition results.

9. A system according to claim 1 , further comprising:

an additional grammar reference module to apply an additional plurality of grammar references to the audio data and to select additional parts of speech most likely to occur in the audio data as additional speech results; and

an additional new grammar module to form an additional new grammar from the additional speech results and to perform speech recognition on the audio data based on the additional new grammar.

10. A method for providing an automated call center inline architecture, comprising:

maintaining prompts and a plurality of grammar references on a script engine;

receiving a call through a telephony interface and collecting audio data using the prompts from the script engine, which are transmitted to the telephony interface via a message server;

performing distributed speech recognition on a speech server, comprising:

receiving the grammar references from the script engine via the message server and determining speech results by applying the grammar references to the audio data;

forming a new grammar from the speech results; and

identifying speech recognition results by applying the new grammar to the audio data;

assigning a confidence level to each of the speech recognition results to provide an indication of accuracy of the distributed speech recognition;

assigning a placeholder to at least a portion of the speech recognition results when the assigned confidence level for those results do not satisfy a predetermined threshold;

performing an action based on a presence of the placeholder, wherein the action comprises querying the message server for additional speech recognition results; and

receiving the speech recognition results and the placeholder as a display on an agent console.

11. A method according to claim 10 , wherein the action further comprises modifying the speech recognition results based on instructions from a user.

12. A method according to claim 10 , further comprising:

simultaneously processing and determining the speech results by a plurality of speech recognizers.

13. A method according to claim 10 , further comprising:

transmitting a text message to the message server upon the receipt of the call through the telephony interface.

14. A method according to claim 10 , wherein the grammar references comprise at least one of independent grammars and non-overlapping sections of a main grammar.

15. A method according to claim 10 , further comprising:

applying an n-best algorithm to the grammar references and the audio data to determine the parts of speech most likely to occur in the audio data.

16. A method according to claim 10 , further comprising:

maintaining a main grammar; and

reforming the main grammar by combining the new grammar with the main grammar.

17. A method according to claim 10 , further comprising:

generating a response to the caller based on the speech recognition results.

18. A method according to claim 10 , further comprising:

applying an additional plurality of grammar references to the audio data;

selecting additional parts of speech most likely to occur in the audio data as additional speech results;

forming an additional new grammar from the additional speech results; and performing speech recognition on the audio data based on the additional new grammar.