vicidial.org

Hey guys,

I'm looking for a soft soundboard (like the one they use to play sound clips)
,its just for telemarketing really. I use ekiga softphones on ubuntu for the client's computer.

I was thinking about coding it myself in PHP using AGI calls or something, but if something is available out of the box, that'd be great and save me a couple of weekends and evenings.

Thank everyone.

You lost me entirely. How does this relate to Vicidial? How do you intend to use this in the call center?

Well, to be honest I hate pitching with a passion. so I sought an easier way to do it with the press of a button.

The point is moot though, because I have found that it was a too awkward way of doing it. Not enough flexibility so half the time I had to pick up the phone anyway.

OK, so you wanted a bunch of buttons with "pitches" or "retorts" to allow you to give stock responses based on the scenario (well-rehearsed, well-spoken retorts for every rejection reason ...) so you could always say it "the right way" without memorization. Next you'll want speech recognition and an AI package to decide which response to use, right?

Hey how did you guess?

I was looking into Sphinx and about to post qs on this exact topic.

Ok thanks. I'm glad that the programming looks fairly simple.

I just tested the script right now, obviously the accuracy is not very good out of the box so I am trying to get this to an acceptable level.

For training, what's the best way to go about it? Right now, I'm looking into doing my own language model just to restrict to the words that are relevant to my script.Will I need to play around the acoustic one as well, or (hopefully) can I leave that alone?

At this point, all I can do is wish you luck and tell you to come back with specific questions ... and remember to Look Around, as there is a bit of movement in the Speech Recog industry right now.

williamconley wrote:OK, so you wanted a bunch of buttons with "pitches" or "retorts" to allow you to give stock responses based on the scenario (well-rehearsed, well-spoken retorts for every rejection reason ...) so you could always say it "the right way" without memorization. Next you'll want speech recognition and an AI package to decide which response to use, right?

Lol funny you mention this I was getting tons of press 1 home security calls from someones vicidial system(Heard the bubble pop sound.) that actually had voice recognition. It was pretty awesome actually. Several times I got transferred to a live rep because I kept playing with it when it called me. It asked how my day was going, etc. If I said bad it said "I am sorry to hear that." and adjusted the opening pitch.

We've had several clients working with various voice recognition systems with varying degrees of success. So far nothing spectacular, but some that are actually pretty good. Unfortunately, I've not had any VR success stories with any of the open source versions. Would be very interesting, though.

We are currently testing an internally developed agent audio soundboard system for use on our VICIhost platform. It has fully integrated web-based administration, and can be built into an agent script or be launched in a separate window when the agent logs in. Currently the soundboard allows an agent to start, stop and restart pre-recorded audio files from the audio store instantly at the click of a mouse. The audio prompts can be arranged in several different ways with main prompts able to have sub-prompts below them.

For more information, go to our vicihost.com website:
http://www.vicihost.com/?p=107

Here's a screenshot:

This is neat. I built something years ago that did something along those lines. When I did it I was bypassing the ACQS by simply having a PHP script triggered from an AJAX call interact with AMI. The production ready method would be to drop something in the vicidial_manager table but maybe i'm confused here. Funny enough it initially started as just something to stream tt-monkeys into agent's meetme confs to mess with their heads, which I found incredibly amusing. It never got as far as having an admin interface built into it but it was a fun AJAX and PHP exercise.

What I did have a need for at one time was a system access a recording based on a unique identifier and than stream it back into the conference so a customer could hear the sale. Kinda like, "Oh, you don't remember signing up?", well here is you saying "sign me up". I got it to work, which required some hoop jumping when the recording was on an archive server because the rec file (sometimes an mp3) would have to traverse the network and placed into tmpfs on the respective telephony server. However I couldn't reliably CONTROL the playback, which is pretty much a deal breaker. Asterisk's ControlPlayback seemed promising but ultimately I think DTMF tones in a meetme were failing. I tried to use a softphone, hardphone and the web interface's send DTMF field but nothing would do it reliably. Maybe in the future a better meetme will be created that will permit me to do what I wanted. Unless someone out there knows a way?

Will this Vicihost soundboard be available for Vicidial? If so, when might it be?

We are still tweaking it, and trying to make it easier to install and use reliably. No plans yet on integration into the vicidial codebase, although the necessary hooks and database fields are already part of the svn/trunk codebase.

Is there any update on this agent soundboard?

Do you have any procedure on how to make this things work?



Vicibox Redux v3.1.14 from iso | Vicidial 2.4-355a BUILD 120118-2113 | Asterisk 1.14.39.2| Single Server | Intel Pentium(R) 4 CPU 3.00 Ghz

It's still only on vicihost for now.

The VICIdial Agent Audio Soundboard is now included in the VICIdial codebase, just upgrade your system, activate the feature in System Settings, and you've got it.

Here's a screenshot of the agent screen with the soundboard loaded:


You can also see what the admin interface for it looks like in our online demo: (6666 / 1234)
http://www.vicidial.org/vicidial_demo/a ... _test&DB=0

HI Matt, i just checked with the latest upgrade of Vicidial but i did not see any settings for Soundboard on how to activate, please can you point us somehow. do we have any Docs on how to used it also..? Thanks

You just enable "Agent Soundboards" in System Settings, then the "Audio Soundboards" menu item will appear in the Admin menus.

I did all that and was able to select or add a new soundboard, then selected an audio, user etc, all saved fine, but when I log in with the soundboard user, nothing shows up on the user side for the soundboard.

Am I missing something.

Ver 2.14.585a Build 170114-1356 All in one server

Ok so I figured it out you have to enable the "script"

<iframe src="./vdc_soundboard_display.php?user=--A--user--B--&pass=--A--pass--B--&bcrypt=OFF&soundboard_id=ComplianceScript" style="background-color:transparent;" scrolling="auto" frameborder="0" allowtransparency="true" width="100%" height="500"> </iframe>

Create a script, (after the soundboard option) put the above in the script, select the script in the campaign and bamm.

Great job.

@chasejordan1

can you give a little step by step process to activate the soundboard.? where do i put the script that you posted, can you share screenshot please, this will help further more. Thanks a lot

@Matt,

Any idea somehow, some steps how to enable the Soundboard for the agents.? any guide how to activate it..please. Thanks a lot

I believe this would qualify as instructions (good post chasejordan1):

chasejordan1 wrote:Create a script, (after the soundboard option) put the below in the script, select the script in the campaign and bamm.

<iframe src="./vdc_soundboard_display.php?user=--A--user--B--&pass=--A--pass--B--&bcrypt=OFF&soundboard_id=ComplianceScript" style="background-color:transparent;" scrolling="auto" frameborder="0" allowtransparency="true" width="100%" height="500"> </iframe>

of course, this would also work as a Web Form.

@williamconley

<iframe src="./vdc_soundboard_display.php?user=--A--user--B--&pass=--A--pass--B--&bcrypt=OFF&soundboard_id=ComplianceScript" style="background-color:transparent;" scrolling="auto" frameborder="0" allowtransparency="true" width="100%" height="500"> </iframe>


i will supply the user/pass and Soundboard ID here right..? in full details.?

@Matt,

Im getting this error somehow upon applying the Soundboard setting on Script, it say invalid user but this user is already a valid user with API set to value 1. can you help somehow please. THanks

Try again with Admin user credentials. If that works, play with the user level and/or compare the Admin user settings to the failing user settings.

I used user 6666 and equivalent password same exact error somehow. What step could i do somehow. Thanks

I'm guessing you have a problem with the script string.

Post what you have verbatim inside a "code block". It shouldn't have anything personal in it anyway, so the whole thing should be safely posted. If you use the full url in the script's src, be sure to leave off your domain/ip, though.

Note that this should NOT contain an actual user and pass. It's expected to have the WORDS "user" and "pass" twice inside this block:

Code: Select all

user=--A--user--B--&pass=--A--pass--B--

These are verbatim. You're not supposed to change them.

A question, if i add a header row, i can only do this by selecting a sound file. Is there a way around this?

You don't have to select a real sound file, you can enter whatever note you want into that field and click submit to just create the entry.

ok thanks.

Also i noticed that if i have 3 columns and the third column has more vertical items then the first two columns the item on the lowest line gets outlined to the left instead of on the horizontal position i specified. Only if i have two adjecent items in the columns left of that 3rd column each item is in the proper column, is that by design?

I wrote it several years ago and I am rarely asked about it, but the last major revision to it was done to customer specs.

@mflorell

Do we have Hotkeys features for vicidial soundboard where agents can generate hotkeys per buttons.? is it possible

That is not currently a feature of agent audio soundboards.