One thing that certainly was new to me this election year was the number of outbound calls we received for political campaigns. They would generally start with "Hi I'm...". I really don't know much more beyond that because by then I had hung up.
In a short excerpt in the Seattle Times these robo calls are described as the lesser of evils when compared to scams and solliciters. I would have to agree to that, as I can hang up very quickly on the outbound call but I have to interrupt the solliciter to take me off their list which, if the solliciter is very insistent, can take me up to a few seconds.
One of these days I may install Speech Server on a box at home to screen all of my calls, but that would be a rather expensive solution with the gateway and all. Of course, a cheaper solution would be to combine TAPI and SAPI in one solution but that would take me more time than I have right now. (Yes, I know such solutions already exist out of the box - but where would the fun be in that?)
Getting back to the point, I was wondering this morning if there were a way to improve those outbound political calls. After all, it isn't a bad idea because it is much cheaper than hiring operators and you have the potential to reach a lot more people.
The first idea I had would be to turn the call into more of a conversation. For instance, instead of Candidate A saying we/he she is the correct choice and why Candidate B is a bad choice why not let the callee ask about a particular issue and then get a response. For instance, what about a call flow similar to the following?
App: Hi, I am John Johnson, running for President. I care about your vote and would like to answer questions you have on my stance on current issues. Do you have any questions? (the key here is this prompt needs to be as short as possible to reduce the chances of people hanging up)
Person: Yes, what is your stance on blah.
App: I think blah is one of the most important problems in our country today. I fully support a plan by nuclear scientist Baga Wada to reduce the blah in our country by 20%. My opponent, on the other hand, recently mispronounced blah in a speech.
To implement this with speech server would just require a well thought out conversational grammar. The app could continue to answer questions until the user is satisfied or hangs up. To add even more value, someone could regularly tune the app and look for failed or low confidence recognitions (because a conversational grammar will never fail - it will always pick one). This person would then make a regular report of the issues not covered by the application (because audio has been recorded) and pass them back to the candidate. The candidate would record his or her answer to the question and the grammar would be modified to contain the answer. Finally, the app could place an outbound call to everyone that asked the question offering them the answer.
To be honest, I didn't stay on the phone long enough to hear whether anyone already does this, but I suspect none of them were that advanced. The system above would still be considered by some to be annoying, but by allowing interaction I think it would be slightly less annoying. Also, given the improvement in recorded prompts and prompt concatenation today, some people may even be fooled into thinking that candidate himself/herself is calling.