Okay, you've got one very basic problem in your logic flow before you start.
I assume this IVR automatically bills the caller each time they press the button. If this is the case, then you don't want the IVR to blurt out it's message until after the caller has identified themselves.
You want this to be as cheap as possible so you can't go making custom phones. You need to work with what the customer already has - a working telephone. What you need is a small black box which does the following.
User presses a button and the box emits a DTMF seuquence which dials the desired number. After a fixed time period the box then emits a second DTMF sequence which identifies the box (customer) for billing purposes.
If this fixed period of time creates a problem, have the user press the same button again after the IVR has answered. The IVR could even prompt the customer to press their button again. This could lead to a few weird issues like the twit holding the box pressing the button out of sequence etc, but you could easily handle that by careful selection of your customer-ID tone sequence ( make sure it is something that won't dial an International Phone number or the like - I'd probably start the sequence with some known tones that the telcos would ignore ) Your IVR would also need to be able to recognise it's own phone number and be able to tell the person to press the button again if they got it wrong. It shouldn't be to hard to come up with a simple scheme which catered to morons who can't even press a button properly, but you had better do that, as they are out there and will probably be using your device. A bit of playing should sort this one out quite easily.
The user picks up their normal phone and holds it normally, they hold your small box against the mouthpiece and press the button. The IVR will answer the call, wait for the Identification tone, play the message and then hang-up. After the user has heard the secret decrypter message they will hang-up their phone too.
You're not going to do this without a battery, but you should be able to make something that will run for years on a single button-cell.
Attempting to power it from the phone line is fraught with problems. For starters you may need telco approval to hang a device off of their phone line. Then you've got the issue of different voltage being used. Contrary to what you may think, it isn't just a nice simple 12v signal. From memory, the ring voltage can be about 90v. Then there is the issue whereby some phones tha tlook like phones aren'r really phones. Well they are, but they don't use anything like a normal phone line as they only work with their own custom PABX or equivelent.
I'm not interested in building this for you. The last time someone wanted to mass produce something they assured me they were going to build in runs of 250,000 of the gadget and wanted it to cost under US$10- After I spent a week doing a feasibility study and told them it could be done, they said that on production of a working proto-type, they would place an order for 10,000 units. That wouldn't have even covered the setup and tooling costs, let alone the purchase of development tools and engineering time. At only 10,000 units all costs would have gone through the roof also. After explaining the simple economics to them, I never heard from them again. Funny that... What really pissed me off is that I had explained all of this to them in detail, before I wasted a week of my time.
Back to your Term Paper. What you need to do is start searching the net for PIC and/or AVR DTMF generators. Good luck with it, it should be pretty simple now that we have defined the task
If you really want to sling a few bucks my way, then send me a PM and let me know how much you are prepared to pay and exactly what has to be done. Be warned though, I am not cheap and you have just used your one free, feasibility study consultation :lol: