So, you wouldn't reccomend Cisco?

We have these new regulations regarding phone payments. I need to be able to get oral authorization before processing and electronic payment. Other than that we use touch tone. Is the cost similar?

No, you misunderstand me.

I was not aware that Cisco have a speech recognition system and was pointing out the partners they do inter-operate with, on the IPIVR (CRS) side, and through MRCP from the voice gateways when using CVP.

Regards,

Geoff

Oh.....I see. So, Cisco doesn't have a speech recognition piece. How does speech recognition occur from a step in the IP-IVR to a different server (Nuance) and then back to the IVR?

I'm a bit rusty on this as I do CVP these days. More knowledgeable posters may chime in here.

In the older versions of CRS, Cisco licenced Nuance or ScanSoft software - it ran on the CRS box. Then when ScanSoft bought Nuance (but decided to use the Nuance name going forward), there was just Nuance.

Then Cisco leveraged the MRCP (Media Resource Control Protocol) from CVP into CRS and now they could support other speech vendors - as they do for CVP.

Regards,

Geoff

Guys, appreciate any other commnents on this topic as I will be loking to learn what is related to VXML, GRXML and its integration to CRS 4.5... Is there anyone out there with knowledge on this? Probably just a flow on how does this work or so...

Thanks very much

Since IPIVR 4.0, Cisco has used the MRCP protocol to communicate with speech rec and TTS engines. The Speech Rec server will need to be separate from the IPIVR server and you will need to purchase the licensing through a Nuance vendor. There are differnet levels of licensing also, simple and complex depending on your speech rec needs. Your Nuance vendor will help you scope this out. Here is a document which outlines the configuration needed on the speech rec server to integrate with IPIVR.

http://www.cisco.com/en/US/prod/collateral/voicesw/custcosw/ps5693/ps1846/prod_bulletin0900aecd806f72f5.html

If you are familiar with CRS scripting, then simple grammars are fairly easy to create. From experience with the Nuance speech rec engine, I know that it does support the GRXML grammars. These are created in the CRS script and passed to the Nuance server at runtime. You can use the built VXML and use the built in voice browser within the CRS scripting, but IPIVR does not fully comply with the VXML 2.0 standard.

Thanks very much for the information, appreciated. Do not why I couldn't find it in first place...