Plum IVR Support

Hi folk,

I have 2 problems:

First problem:
I am testing the French voice (juliette & alain), and it does not work.
Please let me know if is something wrong in my script or if you have some problems with that.
Here the PLUM caller ID:
495179(juliette ) & 495188(alain)

Second Problem:
I tested messages in the answer machine and and I found that one call (ID:495179) took one minute and the other one (ID:495188) took 29s. In both message the system didn't detect the answer machine. Please let me know why the system first,, didn't detect the answer machine, is that normal?, and second and very IMPORTANT thing, why it say that the call took1 minute when that was not real. if you see both calls were the same, only change the voice name.


Regards,
Omar.

Can you provide a sample of your code here rather than simply the IVR outbound call IDs? I'm presuming that this was tested at our IVR hosting environment in the US as those French voices are not available at our European facility.

The IVR outbound dialing system is 95% accurate in determining if a callee is an answering machine or a person. Thus, it does fail from time to time. It should be noted that if an answering machine is initially identified as a person, it is likely that it will almost always be identified as a person. That's because it will consistently answer the phone in the same way such that it confuses the IVR. The overall accuracy of the callee type detection is dependent on a sufficiently large sample size across a diverse range of telephony hardware.

As far as the call lengths are concerned, we'd recommend that you peruse your IVR call logs to look at what happened during each call. The call lengths may differ for a variety of reasons including far-end disconnects over which you have no control. If you would like help determining why two calls might have different call lengths, we will need you to post log excerpts here which we will be able to interpret with you.

Hello,

I am sending the message to:
outbound.plumgroup.com/webservice/queuecall.php
And I am sending that:

POST /webservice/queuecall.php HTTP/1.0
Host: outbound.plumgroup.com
User-Agent: ioncms.com Api
Content-Type: application/x-www-form-urlencoded
Content-Length: 315
login=xxxx
pin=xxxxxxx
phone_number=tel:+xxxxx;ani=xxxxxxx
start_url=http://www.mysite.com/call_start.php
result_url=http://www.mysite.com/call_result.php
message_reference=277
call_parameters=try_confirm=0
max_retries=10
retry_interval=900
scheduled_time=0
expiration_time=


Here is the VXML:
<?xml version="1.0"?>
<vxml version="2.0">
<form>
<block>
<prompt>
<voice name="alain" xml:lang="fr_fr">
Bonjour Omar. Ce message provient du bureau du Dr. Omar Bermudez. Vous avez un rendez-vous avec le Dr. Omar Bermudez le 8 avril à 01:10. Prière de rappeler nos bureaux à 514-xxxxxxx si vous souhaitez reporter votre rendez-vous. Merci. Passez une agréable journée.
</voice>
</prompt>
</block>
</form>
</vxml>

I think that that is all ok, because with English and Spanish voice work well. If you find some error please let me know.


About the detection of the voice or answer machine, I understand that. that is ok. 95% is a good percentage.

About the duration of the call, I don't know what happens, but we are speaking of 100% longer one call from another. That for me is not normal if we know that the messages were the same. Are there some tips or code for to finish (force a disconnect) the call after finish the message?
If not I will check it all my test and will do a report and will show you, because I want to pay for the minutes that I use and I want that my clients pay only for the minutes that they use. May be it was one isolate case, and it is ok, but if not, we will need to see that. For me, that this kind of things are important.

Thank very much
Omar.

Because you are using special language characters in your IVR application, you must specify the encoding that you are using in the XML prolog by adding an "encoding" attribute like this (assuming you're using UTF-8):

<?xml version="1.0" encoding="utf-8"?>

If you don't specify the encoding and the parser comes across a character it doesn't recognize given the default encoding, the speech synthesis will fail.

Please let me know that:

Are there some code for to finish (force a disconnect) the call after finish the message?

Thanks.
Omar.

The IVR tag, <exit/>, should do the trick.

Hello folk,
I am testing Spanish and French voice, and I got the same problem that some days ago, I am sending the vxml, but it didn't work fine, now I added the encoding:

<?xml version="1.0" encoding="utf-8"?>
<vxml version="2.0">
<var name="sessionid" expr="session.id"/>
<var name="confirm" expr="'#N'"/>
<var name="message_reference" expr="'#325'"/>
<form>
<block>
<prompt>
<voice name="alain" xml:lang="fr_fr">
Bonjour Cesar. Ce message provient du bureau du Dr. Agustin Bermudez. Vous avez un rendez-vous avec le Dr. Agustin Bermudez le 5 March �2:10. Pri籥 de rappeler nos bureaux �14-7443603 si vous souhaitez reporter votre rendez-vous. Merci. Passez une agr蠢le journ裮
</voice>
</prompt>
<log>mesref=325</log>
<submit next="http://www.mysite.com/call_confirm.php" namelist="message_reference sessionid confirm" method="post"/>
</block>
</form>
</vxml>

Please could you tell me what is the problem now?
Thanks.
Omar.

Hello,

You should make sure that all of the characters you are sending are also utf-8. Simply adding the encoding at the top of your file does not solve the IVR problem. You should be using a utf-8 capable text editor to create all of your VoiceXML. Also, do not copy and paste from microsoft word. These special characters are not stored as utf-8 in word, you need to type them explictly using the utf-8 compatible editor.

Regards,
Plum Support

I am using French and Spanish characters including in UTF-8 like:
ñ-ç-è-é-à-ô-ó

Let me know if I need to add something more. In my system I open the files with multi-edit or ultra edit and I see the correct character, I write the file with php.

Thanks.
Omar

IVR developers are uncertain that those characters are encoded as UTF-8 with your editor. Please check. They might be ISO-8859-1 or latin-1.

Now YES!!!
the problem was when I save my file, now I save my file in UTF-8 and the French Voice works fine.

Thanks!!